Be the first to post a text review of LAMBDA toolkit. Rate and review a project by clicking thumbs up or thumbs down in the right column.

• LAMBDA version 1.0.0

  I have finally finished incorporating the changes to the bootstrapping functions (see the 6-15-08 news item) into the main distribution and revised the code to the point where I think it is stable enough to call it "version 1.0.0". LAMBDA should now be considered in stable release, and no longer in beta. Thanks to all those who helped and gave me feedback.

  posted by sviscido 485 days ago

• File released: /lambda-source/LAMBDA-1.0.0/LAMBDA_1.0.0.zip

  posted 485 days ago

• File released: /lambda-toolkit/LAMBDA_MRC-1.0.0/LAMBDA_100_MCR_pkg.exe

  posted 485 days ago

• lambda-toolkit LAMBDA_MRC-1.0.0 file released: LAMBDA_100_MCR_pkg.exe

  % README file for the LAMBDA toolkit. % % This is the change log for current and past releases of LAMBDA. Each % release has something new, either major or minor changes to the source % code. Release numbers have three components, x.y.z, such as 1.0.5. The % 'z' refers to a minor release, while 'y' refers to more major changes. % Severe changes (where the whole interface changes completely) would % result in a change to the 'x' number. % % Changes are reported in reverse chronological order (newest listed % first). Note that anything prior to 1.0.0 is inherently considered a % development release, whereas x.9.x is a beta test (0.9.0 is a beta % version of what will, upon full release, be 1.0.0; 1.9.0 would be a beta % version of what will, upon full release, be 0.2.0, etc). Note these are % not decimals, really, so it is possible, for example, to have 1.2.37 (the % 37th minor update to version 1.2). This probably will not happen very % often, of course. % % "New Features" are completely new items that did not exist in a % previous version. "Major Changes" are major alterations to the % behavior of items that previously existed. "Minor changes" are % similar, but for smaller changes. "Fixes" are fixes to things that did % not work. % % PRG = Programmer's Reference Guide % [SV] is Steven Viscido, developer and project post-doc % [EH] is Eli Holmes % [AI] is Anthony Ives (usually, code modified from his MAR scripts) % [SP] is Star-P (changes made as a result of making LAMBDA compatible % with STAR-P software (parallelized matlab) %--------------------------------------------------------------------- % Release 1.0.0 7/25/08 % % Fixes: % * Bootstrapping had not been respecting the user's preferred list of % included/excluded variates and covariates. It now does. % * The above feature caused some conflicts with bootstrapping a single % replicate out of several. This has now been fixed. % % Release 0.9.2 - BETA 7/25/07 % % New Features: % * Excel data can now be read directly into LAMBDA. Make sure your column % headers are in the first row only. (It will also accept data without % column headers). % * You can now display the MAR values (if there are any in memory) to the % command window using a menu item. % % Minor Changes: % * When CIs contain imaginary numbers, a warning dialog now explains why % they are reported as NaNs. % * When importing data, the # of covariates that will be used is shown, % and the user is allowed to say whether to use them or not. % * If you only have one replicate, the plot display no longer asks you how % many reps to show (it shows the one and only one). % % % Fixes: % * Fixed a but that was causing the manual setting of B matrices to give % error messages (it wasn't saving from window to window -- now it is). % * A bug where the best AIC model was being found, but not always correctly % returned, has been fixed. % * Changed some text that displays when the model fit is being assessed. % * Fixed a problem with bootstrapping when C was empty (matlab didn't like % the empty matrix, so it no longer uses C at all when there are no % covariates). %--------------------------------------------------------------------- % Release 0.9.1 - BETA 7/13/07 % % Fixes: % * Fixed a bug in the simulation saving steps where the procedural steps % were not being properly stored/saved. % * Bootstrapped data were not being cleared with other MAR estimates % (despite the fact that the popup help box daid they were). Fixed. %--------------------------------------------------------------------- % Release 0.9.0 - BETA 6/19/07 % % New Features: % * LAMBDA is now invoked as a function call rather than a script. This has % major implications for how data are stored and retrieved (see below). % * Datasets are now partitioned into Classification variables, Variates, % and Covariates (the "Biomass" and "Count" business is now left up to % the user to define, and no longer pre-defined by the program). % * All datasets are now stored in LAMBDA's "work" directory. This should % both help protect against crashes/killing of the program, and also % allows us to avoid either (a) passing massive datasets to functions or % (b) using global variables. % * Since the data format has changed from pre-0.9.0 versions, there is a % "convert" feature that lets one convert old datasets and import them % into the new version. % * SimMAR has now been fully incorporated into LAMBDA. % * It is now possible to re-assign columns of data from one type % (Classification, Variate, Covariate) to another. % * One can now view the data by variable type (Classification, Variate, % Covariate) and thus see just those data, if one wishes. (Use the % 'Data->View' menu.) % * In a dataset that has more than one rep, a set of (contiguous) % replicates can now be deleted. % * LAMBDA is now fully compatible with the matlab runtime compiler, and % can thus be run as a stand-alone. This is a NEW FEATURE so there may be % some issues with it... use at your own risk. % * The user can now take computed means (from the Statistics menu) and put % them back into the main dataset (so you can do CLS on your means % directly, rather than having to export to CVS and re-import). % * The Confidence Limit window now has a new button that will display all % the significant (non-zero) interactions in a nice table in the command % window. This can also be saved to a text file for convenience. % * The significant non-zero B values across multiple reps (using a t-test) % for testing significance) will now be displayed at the user's request % and can be saved to a text file if desired. % % Major Changes: % * The "taxonomy" menu has been removed. The user will need to define % terms and deal with taxonomy externally (it was judged to be outside % the main scope of the program, which is to conduct MAR analysis, not to % act as a database editor). % * SimMAR now exclusively saves and stores files in LAMBDA format, % so we never have to convert anymore (well, other than old SimMAR % files). % % Minor Changes: % * Tweaked the directory structure so that each menu item has its own % folder, besides what other folders have always been present. % * The warning on previewing a dataset (that more than 1000 lines will % be truncated) now only prints if you actually do have > 1000 lines. % * The density (i.e. CPUE) estimation no longer gives a count/hook option, % since that was never implemented anyway. It may be re-added in a future % release if people find they need it. % * Since the workspace is stored as files now, we don't have to worry too % much about whether we need to load it except when starting a CLS % search, so we no longer ask about that when LAMBDA boots up. % * The option to pre-define included variates and covarates has been % removed from the Max Likelihood routines as it is vestigial (it never % really worked and shouldn't -- ML is not done like CLS). % * Stability properties are now all calculated in one click, rather than % requiring several. % * Cleaned up the help text for a lot of the older files. % * The annealing ML function now takes some extra arguments that should % boost performance, making it even faster than it was before. % * SimMAR now warns you if you are working with untransformed covariates, % so that the user will always know (and it stores that information in the % data file). % % % Fixes: % * Removed a few spots where extranneous code information was being % displayed to the command window. % * Preferences are now stored and retrieved correctly every time (in the % past, sometimes LAMBDA was ignoring preferences). % * A couple of spots where "cancel" buttons gave red warning messages have % been fixed. % * Fixed a bug in the delete column function. It now asks for whether you % want to delete variates, covariates, etc, first... and thus only one % type can be deleted at a time. Slightly less convenient, but avoids % data errors. %--------------------------------------------------------------------- % Release 0.7.0 1/31/07 % % New Features: % * A random search may be "seeded" with a known or suspected matrix to try % and improve the search. If existing inclusion matrices are present, the % program will ask whether you want to use it or clear it. This works for % both inlined and non-inlined code, but NOT for Star-P. % % Minor Changes: % * Removed all the previously "extinct" old code from the xtinct folder % (as it has never been needed to date) and moved some "newer" extinct % code into it. % * Saving MAR output now also saves INCLUDED_VARIATES and _COVARS. % % Fixes: % * Fixed a bug where the included variates were not being cleared with the % other MAR variables. % %--------------------------------------------------------------------- % Release 0.6.5 9/31/06 % % New Features: % * There is now a Preferences option (checkbox under edit->preferences) to % allow the "inline" rather than functional CLS algorithm % to be used (this is faster but as it is harder to maintain, it may not be % as 100% up to date as the functional version, as the inlined code is % harder to maintain). [SP, SV] % * There is now a Preferences option (checkbox under edit->preferences) to % tell LAMBDA to use Star-P Parallelization rather than regular % (sequential) iteration. This box should only be checked if you have % Star-P and are running it. If you try it with normal MatLab you will % get error messages. [SP, SV] % % Minor changes: % * Updated the user guide to work with the newest version (0.6.5). % * Made a couple of the dialog boxes smaller than the huge defaults, since % they were mostly wasted space. % * Made some minor tweaks to the syntax to make it more "formally correct" % Matlab (should not change anything from the perspective of the end % user). % % Fixes % * Removed some old code from when LAMBDA (rather than SimMAR) did the % simulations. % * Cleaned up comments in many scripts, and re-compiled Contents.m so it % is completely up to date. %--------------------------------------------------------------------- % Release 0.6.4 8/17/06 % % Minor Changes: % * Simulated Annealing now plots the "temperature graph" only for the % first replicate, if there are multiple reps. % * Re-named a few table entries to be more accurate ("variance" vs. % "stability", and "stability properties" instead of "dynamical % properties"). % * Max lambda B is now displayed when the 'variance' option is chosen (it % was not displayed previously). % * In MAR-1 -> Tools, the 'Check Stability' tool now displays the return % rate and "worst case" reactivity as well as the variance. % * Re-named the ML search box choices (the first one now is "Simplex"). % * Commented out some old debugging steps that have not been used for a % long time (since the debugging is long since finished). This should % improve compute time slightly. % % Fixes: % * ML estimation (simulated annealing version) now correctly deals with % empty C matrices. % * The 'Clear MAR estimates' option now (correctly) clears the arrays of % included variate and covariate information, as well as the ML % estimators (it wasn't before). % * Fixed a minor error with the warning for non-convergence of simplex % (ML) searches. % * Commented out some lines of code in the CLS search that had been used % for error checking but really are not needed in the actual search % algorithm. This should speed up the computation. %--------------------------------------------------------------------- % Release 0.6.3 7/24/06 % % New Features % * There is now a "Compare Ranks" function under MAR-1 -> Tools, that lets % you compare how your CLS or ML estimates ranked interactions, against % ranks from a known B matrix. % * CLS search now keeps track of how long the operation took, and reports % it to the user. % * The mean max(lambda B) is now computed and displayed when other means % are (it's a 2nd way to compute stability, and had been in the MAR % computations but its mean was not being reported along with the rest). % % Minor changes: % * I have gotten rid of the "even/odd" release number scheme. Each release % is stable to my knowledge. % * The CLS search now only displays the FIRST rep's recommended % interactions, rather than all N replicates. (They are all still done -- % just one is displayed to save display space/time.) % * CLS replicate cycling now displays in the command window rather than as % a waitbar. This should slightly speed up the procedure and is also % necessary for Star-P compatibility. [SP] % * Some of the function calls have been moved "up" into higher functions % to speed up the CLS estimation process. [SP] % % Fixes: % * The bootstrapping feature was not storing max Lambda B nor max Lambda % BXB properly. It now is. %--------------------------------------------------------------------- % Release 0.6.2 4/20/06 % % New Features: % * Added a new transformation, the Z-score, for those who wish it (after % Hampton and Schindler 2006, who used it for their MAR analysis). % * MAR averages now strip out any complex values, and report how many % complex values there were in MAR.cplx. % * Added a little MAR tool that lets you check the stability of a known B % matrix (good for testing if your simulation is working, etc). % * Added a WORK.mat file, that is generated each time through an iterated % loop (once per replicate) during the CLS search step -- this should help % with long runs, in case of crashes. You can restore the work from where % you left off. % * The MAR-1 -> Tools menu now has a feature letting the user plot % frequency histograms of data containing several replicates. % % Fixes: % * Fixed a bug where the MAR variable was not properly being globalized. %--------------------------------------------------------------------- % Release 0.6.1 4/20/06 % % New Features: % * Added a new "Load MAR results" feature so you can load the ones you % saved in a previous session. % % Minor changes: % * Modified the FILE menu slightly for readability/usability % * Reduced the font point size on the file name at the bottom of the % window so long file names can be seen. % * Moved the "Clear MAR estimates" option to the bottom of the menu to % prevent mistakes. % %--------------------------------------------------------------------- % Release 0.6.0 3/17/06 % % New Features: % * Added a new "Save MAR results" feature to the file menu to let users % save their MAR-1 analysis for later recall. % * Added the ability to delete a range of rows from the dataset. % % Major Changes: % * The documentation has been completely re-written, and changed to Word % format (PDFs still included) to account for the 6.0 version. % % Minor Changes: % * Made it so that multi-rep MAR only displays the FIRST rep after % assessing the model fits, etc -- which speeds things up. You can still % display them "by hand" with the command line. % % Fixes: % * Removed the "perform QA simulation" menu option, as those functions are % now provided by SimMAR. % * Fixed a bug in the "clear MAR estimates" menu option that was not % actually clearing the whole MAR variable (causing old values to be % incorrectly retained). %--------------------------------------------------------------------- % Release 0.5.9 2/01/06 % % New Features: % * Plotting can now be done of any data, not just simulated data or % means/stds of the stats section. % * Replaced the LAMBDA simulator with a call of the SimMAR toolkit. If you % do not have SimMAR, a warning box pops up and warns you, but no % simulations are done. % * Subsampled data can now be turned into a replicate of the dataset, % allowing users to break one large dataset up into multiple smaller reps % with a few mouse clicks. % * A dataset with multiple replicates can now be merged (the replicates % are "stacked" on top of one another into a single 2-D array). % * Mult-rep computations of A, B, C, and other properties can now have the % mean and STD across reps (mean A, for instance) computed and output to % a file. % % Major Changes: % * Recoded the CLS iteration search to allow for multi-page (multi-rep) % datasets. % * Altered the manual interaction scheme to allow multi-page data. Note % that each page of the interactions is identical here (the user sets it % once and it is multiplied across all reps). % * Altered CLS estimator of the MAR model to loop over, and account for, % multiple reps of data (i.e. more than one dataset). % * Altered ML fminsearch estimators to loop over, and account for, % multiple pages (reps) of data. % * Altered ML annealing estimators to loop over, and account for, multiple % reps. % * Modified the computation of sigma, L, AIC, BIC, etc, to account for % multiple reps. AIC and BIC are now COLUMN vectors, with each rep being % one element of the vector (rather than pages, because it's silly to % have 1 element per page). % * Created a new system, the 'wbpage' scripts and functions, which allows % waitbar generation/updating to take place ONLY if there are multiple % pages (reps) of data. % * Dynamical properties now account for many reps. Note: for properties % that are a single number, each rep is stored as one line of a column % vector; for properties that are a vector or a matrix, each rep is % stored as a "page" (3rd dimension of the array). % * Changed bootstrapping so that it bootstraps for one rep at a time (the % user chooses the rep). Of course, if your data are replicated you % probably don't need to bootstrap, but the option is there if you want. % * Modified data sorting so it will work with multi-rep datasets. % * Column deletion now works for datasets with multiple replicates. % * You can now import multiple reps sequentially from files using the UI. % % Minor Changes: % * Altered the "large dataset" warning of the preview command. % * Modified the "unconcat" function to allow for multi-rep data. % * Re-worked the "Subsample" menu to have groups of operations that are % (hopefully) more logical -- and to accomodate the new "replicate" % feature (see above). % * Moved the global declarations into a separate script for ease. That % script is now executed when LAMBDA starts, rather than each global % being in the lambda.m file. % * Moved the old files that had to do with doing simulations within LAMBDA % to an 'xtinct' folder (spelled that way so it sorts alphabetically to the % bottom). This will be removed in 0.6.0 if everything is Kosher. % * Clearing subsamples now gives a popup window so the user knows it % worked. % * The subsample routine will now give an error and refuse to proceed if % the dataset has more than one replicate. % * Moved the 'clear data' option to the Data menu, rather than File. (It % makes more sense there.) % * Adding columns now checks to see if more than one rep is present and, % if so, reports an error and stops (adding columns across multiple reps is % just too complex to be done easily and not worth coding right now). % * Adjusted the GetData method so it can deal with multiple pages (reps) % now. % * Pooling by rows, taxonomy, and columns can now be done with datasets % that contain multiple replicates. % * Descriptive statistics will ask the user to "merge replicates" into a % single flat array before averaging a multi-rep dataset. This works just % fine. % * When exporting a dataset that has more than one rep to a .csv, LAMBDA % will now force you to merge reps. This will give a usable format % (otherwise reps will be stacked horizontally and won't be identified). % % Fixes: % * Found a spot where sigma was not being recorded properly for univariate % systems. It should be correct now. % * Fixed typos in a couple of menu items % * Removed an unnecessary call to the random number generator in the file % import step (guiload). % * Cancelling text file import will no longer produce an error message. % * Fixed a minor typo in the AIC computation (it wasn't affecting the % actual numbers, but could occasionally generate matlab errors). % * Fixed a typo in the "plot subsample" window (it said "sort" instead of % "plot"). % * Re-namd the 'subsample.m' file because it was conflicting with the name % of the SUBSAMPLE global variable. % * Updated Contents.m using a series of "Contents reports" for the various % subdirectories (and then my own script for concatenating them all). % * Fixed file loading so that if 'cancel' is selected it no longer gives a % MatLab error (cosmetic fix). % * Fixed a bug in GetTaxa where an error was occurring if the 'taxonomy' % field of a data object was empty. %--------------------------------------------------------------------- % Release 0.5.2 10/19/05 % Stable version of 0.5.1. No new features will be added -- only bugfixes % and possibly (very) minor changes. %--------------------------------------------------------------------- % Release 0.5.1 10/19/05 % % Minor changes % * Changed the CLS search step so that it can use either AIC or BIC, at % the operator's discretion (was always AIC before). % * Modified some UI elements slightly to look better (listbox sizes etc). % * There is now a UI-based option for choosing either the Annealing or % Fminsearch methods of doing ML searches. % * Data simulation now has the first covariate as "temperature" going up % a bit every year (to simulate "global warming") rather than being % random (since E is already a random variable). % * Data simulation now simulates the 2nd covariate as "upwelling" and this % is done as a sin wave, though we add +2 to it, to make sure it doesn't go % negative (since we log transform, and log(<0) is not defined). % * Covariates in simulation now will only randomly range from -0.5 to +0.5 % (too large a value was causing models to crash). % * Made it so that clearing data does not clear the simulation parameters. % * Simulation parameters are now clearable separately using the Data menu. % * Added a warning when simulations go below exp(-10), that the parameters % are causing populations to approach extinction. % * Interaction matrix strengths are now set to mean = 0, std = # spp. / 2. % (Example: a 3 spp system has Bs with mean 0 +/- 0.333 STD). The % variance was too high the way I had it before. % % Fixes % * Fixed a typo in the MAR-1 menu % * Fixed a bug in the UI window for data sorting, where it was being % stepped on by another version of one of the variables (one re-used in % another script but not cleared). % * Fixed the automatic and manual "use existing parameters" bit so that it % can only be requested (or automatically tried) if the # of variates and % covariates is the same in the previous and current sim attempts. % * Fixed a bug where if too many imaginary values were computed for % dynamical properties as a result of bootstrapping, an error would result % when computed 95% CIs. Now in such cases, the dynamical property's % confidence limits are reported as NaN and a warning message is % generated. % * Fixed a bug in 'Use Existing Parameters' where only the first Covar was % being used no matter how many there were. %--------------------------------------------------------------------- % Release 0.5.0 10/05/05 % Stable version of 0.5.0. No new features will be added -- only bugfixes % and possibly (very) minor changes. Please note the 0.4.9 comments about % Maximum Likelihood estimates if you are going to try those. % % Minor Changes: % * Updated the Contents.m file (it was out of date). %--------------------------------------------------------------------- % Release 0.4.9 10/03/05 % % New features: % * Added a new UI window that lets the user display the various CI levels % for the different parameters and dynamical properties. MatLab R 13 or % later is required to use this feature. % * Added the UI window listed above as a menu item under MAR-1. % * Added the ability to output all the CI information as a text-file % report. This option is also found under the MAR-1 heading. % * Added the Maximum Likelihood search method. NOTE: This method can be % used after the CLS method to "refine" it or by itself, but if used by % itself, so far it is NOT actually finding the best fit. This still needs % some work. % * You can now clear the MAR estimates (A, B, C, etc) and BOOSTRAP stuff % without having to clear ALL the data (so you can clear and re-run MAR on % the same dataset over and over without re-analyzing or re-loading). % % Minor changes: % * Added a 'use existing parameters' button to the simulation routine so % the user does not have to keep re-inputting the interaction matrix and % other values. % * Added a "scale plots" option to the 2D stat plots so that, if you want, % you can scale the plots to max population size = 1. % * Changed the MAR-1 menu to make it a bit more user-friendly % % Fixes: % * The parameter search was incorrectly forcing a "1" (inclusion) on the % diagonal for COVARIATES as well as variates. It no longer does this. %--------------------------------------------------------------------- % Release 0.4.8 08/25/05 % Stable version of 0.4.7. No new features will be added -- only bugfixes % and possibly (very) minor changes. %--------------------------------------------------------------------- % Release 0.4.7 08/25/05 % % New Features: % * Added a new routine, iteratesearch, which lets the user perform a QA % analysis (Quality Assurance) on the model, UI-accessible. % * Added functions to do QA plots for As and Bs for multiple species, over % model averaging or sample averaging from 1-10 reps. % * Added a "Line Plot" option so the user can "connect the dots" % automatically for 2D stat plots. % * Added a User Guide option to the Help menu, and included a 10-chapter % user's manual in .pdf format. % % Minor changes: % * Added a new preference option for displaying plots (or not) at the end % of a simulation run. % * Modified the subsample pulldowns so they always give you 25 increments % to choose from (giving only a few choices over a small range was not % acceptable). % * References to "CPUE" have been replaced with more generic labels % referring to density or simply "effort" (so that the program is not % fisheries specific). % * Stat plots will now produce error messages if too many categories % or classification variables are used (making plotting impractical). % % Fixes: % * Fixed some typos in this (Readme.m) file. % * Fixed GetData so that if the data object is empty, it returns the null % set rather than giving an error. % * Fixed a bug where the dataset preview was not cutting the display off % at a max of 1,000 rows. % * Fixed a bug in applying the subsample, where it was copying the % contents of the biomass data into (and replacing) the count data. %--------------------------------------------------------------------- % Release 0.4.6 05/27/05 % Stable version of 0.4.5. No new features will be added -- only bugfixes % and possibly (very) minor changes. %--------------------------------------------------------------------- % Release 0.4.5 05/26/05 % % New Features % * Added a new menu item, MAR-1, which contains the UI elements for doing % MAR-1 models with the imported/massaged data. % * A MAR-1 model can now be run from the UI. % * Model assessments can now be run from the UI. % * Dynamical stats (stability, etc) can now be computed from the UI. % * AIC and BIC can now be computed from the UI. % * Bootstrapping and 95% conf limits for bootstrapping can now be done % via the UI. % * Normal probability plotting for each variate is now available from the % UI % * User interface element added to allow users to decide which variates % and covariates will interact with each variate, and which will not. % * User interface element added to do a random search over parameter space % to determine which Bs and Cs to include/exclude. % % Major Changes % * Added a plotting routine at the end of a simulation run to let the user % see graphically what the populations are doing over time. % * Made the plotting routine available at any time after the community has % been simulated. % * Completely re-wrote the (bugged) simulation code so that, with a single % species, 50 years, and no covariates, you can make it EXACTLY do what % Ives' version (in Figure 1) does. % % Minor Changes % * Updated Contents.m to reflect 8 new scripts and some new functions. % * Made it so simulations can go for 5 to 10,000 years (instead of 1-20). % * Added a waitbar to the simulation since doing a lot of years can take a % minute or so... this way the user knows we aren't frozen. % % Fixes % * Fixed a bug in edit->modify->delete where empty data types were % incorrectly being offered as potential columns to be deleted. %--------------------------------------------------------------------- % Release 0.4.4 04/27/05 % Stable version of 0.4.3. No new features will be added -- only bugfixes % and possibly (very) minor changes. %--------------------------------------------------------------------- % Release 0.4.3 patchlevel 0 04/27/05 % Development version including new features. Note - many new functions may % not be accessible from the UI yet. This will change either by the end of % this release or in 0.4.5 and later. % % New Features % * Added a covariate object (oCovar) and modified all the various menu % options so that they will work with it. % * Created functions for computing CLS fitted estimates when given a set % of variates and covariates. Based on the MARBasic script by [AI] % (not working in the UI yet). % * Created a function to find the correlation matrix for the process % error. Based on the MARBasic script by [AI] (not working in the UI % yet). % * Created functions for computing AIC and BIC. Based on the MARBasic % script by [AI] (not working in the UI yet). % % Major Changes % * Modified the various UI elements so they will not "break" if there is % an empty data object (e.g., if you have no Biomass data). % * Modified the Edit Labels window to be cleaner and more in line with the % standard window setup used in the rest of the toolkit. % % Minor Changes % * Removed column-viewing during the label editing step. It was % extranneous. % * Updated Contents.m to reflect the newest coding changes. % * Moved statpolot2d.m into the correct folder (plotfuncs). % * Added 'ztats' class information to Contents.m % % Fixes % * Fixed a bug in pooling data by rows where it always assumed you were % pooling by survey information. % * Fixed the wording in a few dialog boxes ('biomass' was being used where % 'count' was supposed to be in a couple of places due to copy/pasting). %--------------------------------------------------------------------- % Release 0.4.2 03/30/05 % Stable version of 0.4.1. No new features will be added -- only bugfixes % and possibly (very) minor changes. % % Minor Changes % * Began to incorporate the odd-even development-stable format (it was % getting to be too confusing otherwise). Please note that technically % anything before 1.0.0 is still "in development" regardless of oddness % or evenness. But all the "new work" is going on in the latest odd % numbered version from now on. %--------------------------------------------------------------------- % Release 0.4.1 03/30/05 % % New Features % * Created a means of simulating a small community of up to 5 interacting % species, for up to 20 years, at up to 5 stations within the community. % The user can thus put his own A and B parameters in and see what happens % (and then use this to check the model predictions of A and B). % * Added the ability to save, load, and text-export simulation parameters % (i.e. growth rates and interaction terms). % * Updated the PRG to include a nice table of all the callbacks in the % root UI menus. % % Minor Changes % * Updated/corrected Contents.m to reflect new files % * Moved a few files around into the correct directories. % % Fixes % * Tweaked the data_clear script so it works better at clearing globals % (it hadn't been correctly clearing globals, only non-globals). %--------------------------------------------------------------------- % Release 0.3.12 03/28/05 % % New Features: % * Created a new class, ztats, to hold descriptive statistics data (it's % not called "stats" to avoid conflicts with the statistics toolbox). % * Created functionality to do descriptive statistics with up to 5 class % variables (average by year, depth, location, wind speed, etc). % * Created the ability to view the statistics in a table. % * Added stats save, load, and export-to-csv functionality. % * Added the ability to make 3D surface, mesh, and scatter plots of data % with (exactly) 2 class variables. % * Added the ability to make 2D plots of the data % %--------------------------------------------------------------------- % Release 0.3.11 03/15/05 % % New Features: % * Added an 'export to CSV' ability in File --> Export. % * Added a warning box for pre-7.0.1 (R14) users of MatLab. % * Subsample criteria are now recorded in a global variable and can be % saved and viewed by the user. % * Included the ability to add columns to the dataset in memory. % * Included the ability to delete columns from the dataset in memory. % * Included the ability to log-, ln-, sqrt-, and arcsin-transform the % data. % % Minor Changes: % * Re-named the 'Load Raw Data' command to 'Import data', and moved it on % the File menu, so it is more like standard windows positioning. % * Removed relict 'DSET' references (no longer used) in guiload.m. % * Started this change log as the Readme.m file for the LAMBDA toolkit. % * Re-named the root directory of the toolkit to 'LAMBDA' (was 'NRC'). % * Sepcification is requested automatically when a new dataset is loaded, % to mimic behavior of import-data features in programs like Excel. %--------------------------------------------------------------------- % Release 0.3.10 03/09/05 % % New Features: % * Added the ability to pool data by rows for values in a selected column % to Statistics --> Pool. % % Minor Changes: % * Created several functions to replace repetitive code in a few places. % %

  posted 485 days ago

• lambda-source LAMBDA-1.0.0 file released: LAMBDA_1.0.0.zip

  % README file for the LAMBDA toolkit. % % This is the change log for current and past releases of LAMBDA. Each % release has something new, either major or minor changes to the source % code. Release numbers have three components, x.y.z, such as 1.0.5. The % 'z' refers to a minor release, while 'y' refers to more major changes. % Severe changes (where the whole interface changes completely) would % result in a change to the 'x' number. % % Changes are reported in reverse chronological order (newest listed % first). Note that anything prior to 1.0.0 is inherently considered a % development release, whereas x.9.x is a beta test (0.9.0 is a beta % version of what will, upon full release, be 1.0.0; 1.9.0 would be a beta % version of what will, upon full release, be 0.2.0, etc). Note these are % not decimals, really, so it is possible, for example, to have 1.2.37 (the % 37th minor update to version 1.2). This probably will not happen very % often, of course. % % "New Features" are completely new items that did not exist in a % previous version. "Major Changes" are major alterations to the % behavior of items that previously existed. "Minor changes" are % similar, but for smaller changes. "Fixes" are fixes to things that did % not work. % % PRG = Programmer's Reference Guide % [SV] is Steven Viscido, developer and project post-doc % [EH] is Eli Holmes % [AI] is Anthony Ives (usually, code modified from his MAR scripts) % [SP] is Star-P (changes made as a result of making LAMBDA compatible % with STAR-P software (parallelized matlab) %--------------------------------------------------------------------- % Release 1.0.0 7/25/08 % % Fixes: % * Bootstrapping had not been respecting the user's preferred list of % included/excluded variates and covariates. It now does. % * The above feature caused some conflicts with bootstrapping a single % replicate out of several. This has now been fixed. % % Release 0.9.2 - BETA 7/25/07 % % New Features: % * Excel data can now be read directly into LAMBDA. Make sure your column % headers are in the first row only. (It will also accept data without % column headers). % * You can now display the MAR values (if there are any in memory) to the % command window using a menu item. % % Minor Changes: % * When CIs contain imaginary numbers, a warning dialog now explains why % they are reported as NaNs. % * When importing data, the # of covariates that will be used is shown, % and the user is allowed to say whether to use them or not. % * If you only have one replicate, the plot display no longer asks you how % many reps to show (it shows the one and only one). % % % Fixes: % * Fixed a but that was causing the manual setting of B matrices to give % error messages (it wasn't saving from window to window -- now it is). % * A bug where the best AIC model was being found, but not always correctly % returned, has been fixed. % * Changed some text that displays when the model fit is being assessed. % * Fixed a problem with bootstrapping when C was empty (matlab didn't like % the empty matrix, so it no longer uses C at all when there are no % covariates). %--------------------------------------------------------------------- % Release 0.9.1 - BETA 7/13/07 % % Fixes: % * Fixed a bug in the simulation saving steps where the procedural steps % were not being properly stored/saved. % * Bootstrapped data were not being cleared with other MAR estimates % (despite the fact that the popup help box daid they were). Fixed. %--------------------------------------------------------------------- % Release 0.9.0 - BETA 6/19/07 % % New Features: % * LAMBDA is now invoked as a function call rather than a script. This has % major implications for how data are stored and retrieved (see below). % * Datasets are now partitioned into Classification variables, Variates, % and Covariates (the "Biomass" and "Count" business is now left up to % the user to define, and no longer pre-defined by the program). % * All datasets are now stored in LAMBDA's "work" directory. This should % both help protect against crashes/killing of the program, and also % allows us to avoid either (a) passing massive datasets to functions or % (b) using global variables. % * Since the data format has changed from pre-0.9.0 versions, there is a % "convert" feature that lets one convert old datasets and import them % into the new version. % * SimMAR has now been fully incorporated into LAMBDA. % * It is now possible to re-assign columns of data from one type % (Classification, Variate, Covariate) to another. % * One can now view the data by variable type (Classification, Variate, % Covariate) and thus see just those data, if one wishes. (Use the % 'Data->View' menu.) % * In a dataset that has more than one rep, a set of (contiguous) % replicates can now be deleted. % * LAMBDA is now fully compatible with the matlab runtime compiler, and % can thus be run as a stand-alone. This is a NEW FEATURE so there may be % some issues with it... use at your own risk. % * The user can now take computed means (from the Statistics menu) and put % them back into the main dataset (so you can do CLS on your means % directly, rather than having to export to CVS and re-import). % * The Confidence Limit window now has a new button that will display all % the significant (non-zero) interactions in a nice table in the command % window. This can also be saved to a text file for convenience. % * The significant non-zero B values across multiple reps (using a t-test) % for testing significance) will now be displayed at the user's request % and can be saved to a text file if desired. % % Major Changes: % * The "taxonomy" menu has been removed. The user will need to define % terms and deal with taxonomy externally (it was judged to be outside % the main scope of the program, which is to conduct MAR analysis, not to % act as a database editor). % * SimMAR now exclusively saves and stores files in LAMBDA format, % so we never have to convert anymore (well, other than old SimMAR % files). % % Minor Changes: % * Tweaked the directory structure so that each menu item has its own % folder, besides what other folders have always been present. % * The warning on previewing a dataset (that more than 1000 lines will % be truncated) now only prints if you actually do have > 1000 lines. % * The density (i.e. CPUE) estimation no longer gives a count/hook option, % since that was never implemented anyway. It may be re-added in a future % release if people find they need it. % * Since the workspace is stored as files now, we don't have to worry too % much about whether we need to load it except when starting a CLS % search, so we no longer ask about that when LAMBDA boots up. % * The option to pre-define included variates and covarates has been % removed from the Max Likelihood routines as it is vestigial (it never % really worked and shouldn't -- ML is not done like CLS). % * Stability properties are now all calculated in one click, rather than % requiring several. % * Cleaned up the help text for a lot of the older files. % * The annealing ML function now takes some extra arguments that should % boost performance, making it even faster than it was before. % * SimMAR now warns you if you are working with untransformed covariates, % so that the user will always know (and it stores that information in the % data file). % % % Fixes: % * Removed a few spots where extranneous code information was being % displayed to the command window. % * Preferences are now stored and retrieved correctly every time (in the % past, sometimes LAMBDA was ignoring preferences). % * A couple of spots where "cancel" buttons gave red warning messages have % been fixed. % * Fixed a bug in the delete column function. It now asks for whether you % want to delete variates, covariates, etc, first... and thus only one % type can be deleted at a time. Slightly less convenient, but avoids % data errors. %--------------------------------------------------------------------- % Release 0.7.0 1/31/07 % % New Features: % * A random search may be "seeded" with a known or suspected matrix to try % and improve the search. If existing inclusion matrices are present, the % program will ask whether you want to use it or clear it. This works for % both inlined and non-inlined code, but NOT for Star-P. % % Minor Changes: % * Removed all the previously "extinct" old code from the xtinct folder % (as it has never been needed to date) and moved some "newer" extinct % code into it. % * Saving MAR output now also saves INCLUDED_VARIATES and _COVARS. % % Fixes: % * Fixed a bug where the included variates were not being cleared with the % other MAR variables. % %--------------------------------------------------------------------- % Release 0.6.5 9/31/06 % % New Features: % * There is now a Preferences option (checkbox under edit->preferences) to % allow the "inline" rather than functional CLS algorithm % to be used (this is faster but as it is harder to maintain, it may not be % as 100% up to date as the functional version, as the inlined code is % harder to maintain). [SP, SV] % * There is now a Preferences option (checkbox under edit->preferences) to % tell LAMBDA to use Star-P Parallelization rather than regular % (sequential) iteration. This box should only be checked if you have % Star-P and are running it. If you try it with normal MatLab you will % get error messages. [SP, SV] % % Minor changes: % * Updated the user guide to work with the newest version (0.6.5). % * Made a couple of the dialog boxes smaller than the huge defaults, since % they were mostly wasted space. % * Made some minor tweaks to the syntax to make it more "formally correct" % Matlab (should not change anything from the perspective of the end % user). % % Fixes % * Removed some old code from when LAMBDA (rather than SimMAR) did the % simulations. % * Cleaned up comments in many scripts, and re-compiled Contents.m so it % is completely up to date. %--------------------------------------------------------------------- % Release 0.6.4 8/17/06 % % Minor Changes: % * Simulated Annealing now plots the "temperature graph" only for the % first replicate, if there are multiple reps. % * Re-named a few table entries to be more accurate ("variance" vs. % "stability", and "stability properties" instead of "dynamical % properties"). % * Max lambda B is now displayed when the 'variance' option is chosen (it % was not displayed previously). % * In MAR-1 -> Tools, the 'Check Stability' tool now displays the return % rate and "worst case" reactivity as well as the variance. % * Re-named the ML search box choices (the first one now is "Simplex"). % * Commented out some old debugging steps that have not been used for a % long time (since the debugging is long since finished). This should % improve compute time slightly. % % Fixes: % * ML estimation (simulated annealing version) now correctly deals with % empty C matrices. % * The 'Clear MAR estimates' option now (correctly) clears the arrays of % included variate and covariate information, as well as the ML % estimators (it wasn't before). % * Fixed a minor error with the warning for non-convergence of simplex % (ML) searches. % * Commented out some lines of code in the CLS search that had been used % for error checking but really are not needed in the actual search % algorithm. This should speed up the computation. %--------------------------------------------------------------------- % Release 0.6.3 7/24/06 % % New Features % * There is now a "Compare Ranks" function under MAR-1 -> Tools, that lets % you compare how your CLS or ML estimates ranked interactions, against % ranks from a known B matrix. % * CLS search now keeps track of how long the operation took, and reports % it to the user. % * The mean max(lambda B) is now computed and displayed when other means % are (it's a 2nd way to compute stability, and had been in the MAR % computations but its mean was not being reported along with the rest). % % Minor changes: % * I have gotten rid of the "even/odd" release number scheme. Each release % is stable to my knowledge. % * The CLS search now only displays the FIRST rep's recommended % interactions, rather than all N replicates. (They are all still done -- % just one is displayed to save display space/time.) % * CLS replicate cycling now displays in the command window rather than as % a waitbar. This should slightly speed up the procedure and is also % necessary for Star-P compatibility. [SP] % * Some of the function calls have been moved "up" into higher functions % to speed up the CLS estimation process. [SP] % % Fixes: % * The bootstrapping feature was not storing max Lambda B nor max Lambda % BXB properly. It now is. %--------------------------------------------------------------------- % Release 0.6.2 4/20/06 % % New Features: % * Added a new transformation, the Z-score, for those who wish it (after % Hampton and Schindler 2006, who used it for their MAR analysis). % * MAR averages now strip out any complex values, and report how many % complex values there were in MAR.cplx. % * Added a little MAR tool that lets you check the stability of a known B % matrix (good for testing if your simulation is working, etc). % * Added a WORK.mat file, that is generated each time through an iterated % loop (once per replicate) during the CLS search step -- this should help % with long runs, in case of crashes. You can restore the work from where % you left off. % * The MAR-1 -> Tools menu now has a feature letting the user plot % frequency histograms of data containing several replicates. % % Fixes: % * Fixed a bug where the MAR variable was not properly being globalized. %--------------------------------------------------------------------- % Release 0.6.1 4/20/06 % % New Features: % * Added a new "Load MAR results" feature so you can load the ones you % saved in a previous session. % % Minor changes: % * Modified the FILE menu slightly for readability/usability % * Reduced the font point size on the file name at the bottom of the % window so long file names can be seen. % * Moved the "Clear MAR estimates" option to the bottom of the menu to % prevent mistakes. % %--------------------------------------------------------------------- % Release 0.6.0 3/17/06 % % New Features: % * Added a new "Save MAR results" feature to the file menu to let users % save their MAR-1 analysis for later recall. % * Added the ability to delete a range of rows from the dataset. % % Major Changes: % * The documentation has been completely re-written, and changed to Word % format (PDFs still included) to account for the 6.0 version. % % Minor Changes: % * Made it so that multi-rep MAR only displays the FIRST rep after % assessing the model fits, etc -- which speeds things up. You can still % display them "by hand" with the command line. % % Fixes: % * Removed the "perform QA simulation" menu option, as those functions are % now provided by SimMAR. % * Fixed a bug in the "clear MAR estimates" menu option that was not % actually clearing the whole MAR variable (causing old values to be % incorrectly retained). %--------------------------------------------------------------------- % Release 0.5.9 2/01/06 % % New Features: % * Plotting can now be done of any data, not just simulated data or % means/stds of the stats section. % * Replaced the LAMBDA simulator with a call of the SimMAR toolkit. If you % do not have SimMAR, a warning box pops up and warns you, but no % simulations are done. % * Subsampled data can now be turned into a replicate of the dataset, % allowing users to break one large dataset up into multiple smaller reps % with a few mouse clicks. % * A dataset with multiple replicates can now be merged (the replicates % are "stacked" on top of one another into a single 2-D array). % * Mult-rep computations of A, B, C, and other properties can now have the % mean and STD across reps (mean A, for instance) computed and output to % a file. % % Major Changes: % * Recoded the CLS iteration search to allow for multi-page (multi-rep) % datasets. % * Altered the manual interaction scheme to allow multi-page data. Note % that each page of the interactions is identical here (the user sets it % once and it is multiplied across all reps). % * Altered CLS estimator of the MAR model to loop over, and account for, % multiple reps of data (i.e. more than one dataset). % * Altered ML fminsearch estimators to loop over, and account for, % multiple pages (reps) of data. % * Altered ML annealing estimators to loop over, and account for, multiple % reps. % * Modified the computation of sigma, L, AIC, BIC, etc, to account for % multiple reps. AIC and BIC are now COLUMN vectors, with each rep being % one element of the vector (rather than pages, because it's silly to % have 1 element per page). % * Created a new system, the 'wbpage' scripts and functions, which allows % waitbar generation/updating to take place ONLY if there are multiple % pages (reps) of data. % * Dynamical properties now account for many reps. Note: for properties % that are a single number, each rep is stored as one line of a column % vector; for properties that are a vector or a matrix, each rep is % stored as a "page" (3rd dimension of the array). % * Changed bootstrapping so that it bootstraps for one rep at a time (the % user chooses the rep). Of course, if your data are replicated you % probably don't need to bootstrap, but the option is there if you want. % * Modified data sorting so it will work with multi-rep datasets. % * Column deletion now works for datasets with multiple replicates. % * You can now import multiple reps sequentially from files using the UI. % % Minor Changes: % * Altered the "large dataset" warning of the preview command. % * Modified the "unconcat" function to allow for multi-rep data. % * Re-worked the "Subsample" menu to have groups of operations that are % (hopefully) more logical -- and to accomodate the new "replicate" % feature (see above). % * Moved the global declarations into a separate script for ease. That % script is now executed when LAMBDA starts, rather than each global % being in the lambda.m file. % * Moved the old files that had to do with doing simulations within LAMBDA % to an 'xtinct' folder (spelled that way so it sorts alphabetically to the % bottom). This will be removed in 0.6.0 if everything is Kosher. % * Clearing subsamples now gives a popup window so the user knows it % worked. % * The subsample routine will now give an error and refuse to proceed if % the dataset has more than one replicate. % * Moved the 'clear data' option to the Data menu, rather than File. (It % makes more sense there.) % * Adding columns now checks to see if more than one rep is present and, % if so, reports an error and stops (adding columns across multiple reps is % just too complex to be done easily and not worth coding right now). % * Adjusted the GetData method so it can deal with multiple pages (reps) % now. % * Pooling by rows, taxonomy, and columns can now be done with datasets % that contain multiple replicates. % * Descriptive statistics will ask the user to "merge replicates" into a % single flat array before averaging a multi-rep dataset. This works just % fine. % * When exporting a dataset that has more than one rep to a .csv, LAMBDA % will now force you to merge reps. This will give a usable format % (otherwise reps will be stacked horizontally and won't be identified). % % Fixes: % * Found a spot where sigma was not being recorded properly for univariate % systems. It should be correct now. % * Fixed typos in a couple of menu items % * Removed an unnecessary call to the random number generator in the file % import step (guiload). % * Cancelling text file import will no longer produce an error message. % * Fixed a minor typo in the AIC computation (it wasn't affecting the % actual numbers, but could occasionally generate matlab errors). % * Fixed a typo in the "plot subsample" window (it said "sort" instead of % "plot"). % * Re-namd the 'subsample.m' file because it was conflicting with the name % of the SUBSAMPLE global variable. % * Updated Contents.m using a series of "Contents reports" for the various % subdirectories (and then my own script for concatenating them all). % * Fixed file loading so that if 'cancel' is selected it no longer gives a % MatLab error (cosmetic fix). % * Fixed a bug in GetTaxa where an error was occurring if the 'taxonomy' % field of a data object was empty. %--------------------------------------------------------------------- % Release 0.5.2 10/19/05 % Stable version of 0.5.1. No new features will be added -- only bugfixes % and possibly (very) minor changes. %--------------------------------------------------------------------- % Release 0.5.1 10/19/05 % % Minor changes % * Changed the CLS search step so that it can use either AIC or BIC, at % the operator's discretion (was always AIC before). % * Modified some UI elements slightly to look better (listbox sizes etc). % * There is now a UI-based option for choosing either the Annealing or % Fminsearch methods of doing ML searches. % * Data simulation now has the first covariate as "temperature" going up % a bit every year (to simulate "global warming") rather than being % random (since E is already a random variable). % * Data simulation now simulates the 2nd covariate as "upwelling" and this % is done as a sin wave, though we add +2 to it, to make sure it doesn't go % negative (since we log transform, and log(<0) is not defined). % * Covariates in simulation now will only randomly range from -0.5 to +0.5 % (too large a value was causing models to crash). % * Made it so that clearing data does not clear the simulation parameters. % * Simulation parameters are now clearable separately using the Data menu. % * Added a warning when simulations go below exp(-10), that the parameters % are causing populations to approach extinction. % * Interaction matrix strengths are now set to mean = 0, std = # spp. / 2. % (Example: a 3 spp system has Bs with mean 0 +/- 0.333 STD). The % variance was too high the way I had it before. % % Fixes % * Fixed a typo in the MAR-1 menu % * Fixed a bug in the UI window for data sorting, where it was being % stepped on by another version of one of the variables (one re-used in % another script but not cleared). % * Fixed the automatic and manual "use existing parameters" bit so that it % can only be requested (or automatically tried) if the # of variates and % covariates is the same in the previous and current sim attempts. % * Fixed a bug where if too many imaginary values were computed for % dynamical properties as a result of bootstrapping, an error would result % when computed 95% CIs. Now in such cases, the dynamical property's % confidence limits are reported as NaN and a warning message is % generated. % * Fixed a bug in 'Use Existing Parameters' where only the first Covar was % being used no matter how many there were. %--------------------------------------------------------------------- % Release 0.5.0 10/05/05 % Stable version of 0.5.0. No new features will be added -- only bugfixes % and possibly (very) minor changes. Please note the 0.4.9 comments about % Maximum Likelihood estimates if you are going to try those. % % Minor Changes: % * Updated the Contents.m file (it was out of date). %--------------------------------------------------------------------- % Release 0.4.9 10/03/05 % % New features: % * Added a new UI window that lets the user display the various CI levels % for the different parameters and dynamical properties. MatLab R 13 or % later is required to use this feature. % * Added the UI window listed above as a menu item under MAR-1. % * Added the ability to output all the CI information as a text-file % report. This option is also found under the MAR-1 heading. % * Added the Maximum Likelihood search method. NOTE: This method can be % used after the CLS method to "refine" it or by itself, but if used by % itself, so far it is NOT actually finding the best fit. This still needs % some work. % * You can now clear the MAR estimates (A, B, C, etc) and BOOSTRAP stuff % without having to clear ALL the data (so you can clear and re-run MAR on % the same dataset over and over without re-analyzing or re-loading). % % Minor changes: % * Added a 'use existing parameters' button to the simulation routine so % the user does not have to keep re-inputting the interaction matrix and % other values. % * Added a "scale plots" option to the 2D stat plots so that, if you want, % you can scale the plots to max population size = 1. % * Changed the MAR-1 menu to make it a bit more user-friendly % % Fixes: % * The parameter search was incorrectly forcing a "1" (inclusion) on the % diagonal for COVARIATES as well as variates. It no longer does this. %--------------------------------------------------------------------- % Release 0.4.8 08/25/05 % Stable version of 0.4.7. No new features will be added -- only bugfixes % and possibly (very) minor changes. %--------------------------------------------------------------------- % Release 0.4.7 08/25/05 % % New Features: % * Added a new routine, iteratesearch, which lets the user perform a QA % analysis (Quality Assurance) on the model, UI-accessible. % * Added functions to do QA plots for As and Bs for multiple species, over % model averaging or sample averaging from 1-10 reps. % * Added a "Line Plot" option so the user can "connect the dots" % automatically for 2D stat plots. % * Added a User Guide option to the Help menu, and included a 10-chapter % user's manual in .pdf format. % % Minor changes: % * Added a new preference option for displaying plots (or not) at the end % of a simulation run. % * Modified the subsample pulldowns so they always give you 25 increments % to choose from (giving only a few choices over a small range was not % acceptable). % * References to "CPUE" have been replaced with more generic labels % referring to density or simply "effort" (so that the program is not % fisheries specific). % * Stat plots will now produce error messages if too many categories % or classification variables are used (making plotting impractical). % % Fixes: % * Fixed some typos in this (Readme.m) file. % * Fixed GetData so that if the data object is empty, it returns the null % set rather than giving an error. % * Fixed a bug where the dataset preview was not cutting the display off % at a max of 1,000 rows. % * Fixed a bug in applying the subsample, where it was copying the % contents of the biomass data into (and replacing) the count data. %--------------------------------------------------------------------- % Release 0.4.6 05/27/05 % Stable version of 0.4.5. No new features will be added -- only bugfixes % and possibly (very) minor changes. %--------------------------------------------------------------------- % Release 0.4.5 05/26/05 % % New Features % * Added a new menu item, MAR-1, which contains the UI elements for doing % MAR-1 models with the imported/massaged data. % * A MAR-1 model can now be run from the UI. % * Model assessments can now be run from the UI. % * Dynamical stats (stability, etc) can now be computed from the UI. % * AIC and BIC can now be computed from the UI. % * Bootstrapping and 95% conf limits for bootstrapping can now be done % via the UI. % * Normal probability plotting for each variate is now available from the % UI % * User interface element added to allow users to decide which variates % and covariates will interact with each variate, and which will not. % * User interface element added to do a random search over parameter space % to determine which Bs and Cs to include/exclude. % % Major Changes % * Added a plotting routine at the end of a simulation run to let the user % see graphically what the populations are doing over time. % * Made the plotting routine available at any time after the community has % been simulated. % * Completely re-wrote the (bugged) simulation code so that, with a single % species, 50 years, and no covariates, you can make it EXACTLY do what % Ives' version (in Figure 1) does. % % Minor Changes % * Updated Contents.m to reflect 8 new scripts and some new functions. % * Made it so simulations can go for 5 to 10,000 years (instead of 1-20). % * Added a waitbar to the simulation since doing a lot of years can take a % minute or so... this way the user knows we aren't frozen. % % Fixes % * Fixed a bug in edit->modify->delete where empty data types were % incorrectly being offered as potential columns to be deleted. %--------------------------------------------------------------------- % Release 0.4.4 04/27/05 % Stable version of 0.4.3. No new features will be added -- only bugfixes % and possibly (very) minor changes. %--------------------------------------------------------------------- % Release 0.4.3 patchlevel 0 04/27/05 % Development version including new features. Note - many new functions may % not be accessible from the UI yet. This will change either by the end of % this release or in 0.4.5 and later. % % New Features % * Added a covariate object (oCovar) and modified all the various menu % options so that they will work with it. % * Created functions for computing CLS fitted estimates when given a set % of variates and covariates. Based on the MARBasic script by [AI] % (not working in the UI yet). % * Created a function to find the correlation matrix for the process % error. Based on the MARBasic script by [AI] (not working in the UI % yet). % * Created functions for computing AIC and BIC. Based on the MARBasic % script by [AI] (not working in the UI yet). % % Major Changes % * Modified the various UI elements so they will not "break" if there is % an empty data object (e.g., if you have no Biomass data). % * Modified the Edit Labels window to be cleaner and more in line with the % standard window setup used in the rest of the toolkit. % % Minor Changes % * Removed column-viewing during the label editing step. It was % extranneous. % * Updated Contents.m to reflect the newest coding changes. % * Moved statpolot2d.m into the correct folder (plotfuncs). % * Added 'ztats' class information to Contents.m % % Fixes % * Fixed a bug in pooling data by rows where it always assumed you were % pooling by survey information. % * Fixed the wording in a few dialog boxes ('biomass' was being used where % 'count' was supposed to be in a couple of places due to copy/pasting). %--------------------------------------------------------------------- % Release 0.4.2 03/30/05 % Stable version of 0.4.1. No new features will be added -- only bugfixes % and possibly (very) minor changes. % % Minor Changes % * Began to incorporate the odd-even development-stable format (it was % getting to be too confusing otherwise). Please note that technically % anything before 1.0.0 is still "in development" regardless of oddness % or evenness. But all the "new work" is going on in the latest odd % numbered version from now on. %--------------------------------------------------------------------- % Release 0.4.1 03/30/05 % % New Features % * Created a means of simulating a small community of up to 5 interacting % species, for up to 20 years, at up to 5 stations within the community. % The user can thus put his own A and B parameters in and see what happens % (and then use this to check the model predictions of A and B). % * Added the ability to save, load, and text-export simulation parameters % (i.e. growth rates and interaction terms). % * Updated the PRG to include a nice table of all the callbacks in the % root UI menus. % % Minor Changes % * Updated/corrected Contents.m to reflect new files % * Moved a few files around into the correct directories. % % Fixes % * Tweaked the data_clear script so it works better at clearing globals % (it hadn't been correctly clearing globals, only non-globals). %--------------------------------------------------------------------- % Release 0.3.12 03/28/05 % % New Features: % * Created a new class, ztats, to hold descriptive statistics data (it's % not called "stats" to avoid conflicts with the statistics toolbox). % * Created functionality to do descriptive statistics with up to 5 class % variables (average by year, depth, location, wind speed, etc). % * Created the ability to view the statistics in a table. % * Added stats save, load, and export-to-csv functionality. % * Added the ability to make 3D surface, mesh, and scatter plots of data % with (exactly) 2 class variables. % * Added the ability to make 2D plots of the data % %--------------------------------------------------------------------- % Release 0.3.11 03/15/05 % % New Features: % * Added an 'export to CSV' ability in File --> Export. % * Added a warning box for pre-7.0.1 (R14) users of MatLab. % * Subsample criteria are now recorded in a global variable and can be % saved and viewed by the user. % * Included the ability to add columns to the dataset in memory. % * Included the ability to delete columns from the dataset in memory. % * Included the ability to log-, ln-, sqrt-, and arcsin-transform the % data. % % Minor Changes: % * Re-named the 'Load Raw Data' command to 'Import data', and moved it on % the File menu, so it is more like standard windows positioning. % * Removed relict 'DSET' references (no longer used) in guiload.m. % * Started this change log as the Readme.m file for the LAMBDA toolkit. % * Re-named the root directory of the toolkit to 'LAMBDA' (was 'NRC'). % * Sepcification is requested automatically when a new dataset is loaded, % to mimic behavior of import-data features in programs like Excel. %--------------------------------------------------------------------- % Release 0.3.10 03/09/05 % % New Features: % * Added the ability to pool data by rows for values in a selected column % to Statistics --> Pool. % % Minor Changes: % * Created several functions to replace repetitive code in a few places. % %

  posted 485 days ago

• Bootstrap function updated 06-15-2008

  There was a slight issue with the bootstrapping function. I have not had time to thoroughly test the fix, so I have not made a new version yet. However, the file to fix the problem is located in the &quot;all files&quot; section, called DoBootstrap.m. If you download this file and place it into the MAR folder in lambda's main folder, this will correct the problem -- which was that it was not &quot;respecting&quot; your choices if you were telling it to definitely include or exclude variates or covariates... it was just trying them all. When I get this more fully tested I will merge it into the main source tree and post it as version 0.9.3. Steve

  posted by sviscido 529 days ago

• File released: /lambda-source/0.9.2-Beta/DoBootstraps.m

  posted 529 days ago

• lambda-source 0.9.2-Beta file released: DoBootstraps.m

  posted 529 days ago

• LAMBDA Wiki completed!

  The wiki-fication of the LAMBDA documentation is now basically done. I need to eventually add some screenshots for illustration, but all the text of the PDF and Word file user guides has been imported into the wiki. Thus, LAMBDA documentation can now be viewed online directly as well as downloaded.

  posted by sviscido 645 days ago

• LAMBDA Wiki

  I have begun slowly incorporating the text of the latest draft of the LAMBDA documentation (which is kept in Word and PDF files) into the Wiki section here at SourceForge. The goal is to eventually have the entire document essentially &quot;mirrored&quot; here at SourceForge on the Wiki, so that there is a way to do online searches for documentation items. Also, once it's done I will allow other people to update it, which will hopefully make this a vibrant and dynamic, and more thorough, documentation of the software.

  posted by sviscido 649 days ago