Re: [ojAlgo-user] Getting started with ojAlgo

[Anders P. <an...@op...>]

On 2 nov 2009, at 15.22, Alok Saldanha wrote:

> I am having trouble getting started with ojAlgo. I am confused by  
> the class structure and am not sure how to do basic things. Are  
> there tutorials anywhere?


None that I know of.


> I would like to be able to do the following
>
> 1) Load historical data for two tech stocks and the nasdaq that I  
> have already downloaded from yahoo
> 2) Calculate raw covariance matrix for the two tech stocks.
> 3) Solve a MarkowitzModel using some expected excess return values  
> also stored on disk.
>
> In order to load historical data, I modified SymbolData so that  
> fetch could take a BufferedReader argument that I built off a file,
>
>     public void fetch() {
>         BufferedReader tmpInput = myResourceLocator.getStreamReader();
>         fetch(tmpInput);
>     }
>     public void fetch(BufferedReader tmpInput) {
>        ... existing code here...
>     }
>
> It is somewhat surprising to me that I had to modify the library in  
> order to load data off disk rather than over the network, but I  
> couldn't figure out an easier way.


The code in the org.ojalgo.finance.data package is designed to load  
data from Yahoo or Google - that's what it does. The use case you  
describe never occurred to me. (I never store data in Yahoo/Google  
format on file.)


> I have not been able to figure out how to calculate covariance; I am  
> able to populate a matrix with my data values as follows (nokPrices,  
> orclPrices and nasdaqPrices are of type List<HistoricalData>, which  
> I made public so I could call SymbolData.getHistoricalPrices()):


What I do is call/use SymbolDataCache.getInstance().getPriceSeries(...)

I'll think about making the changes you suggested.


>         MatrixStore<Double> a =  PrimitiveDenseStore.FACTORY.makeZero 
> (3,
>                 nokPrices.size());
>         BasicMatrix b = new PrimitiveMatrix(a);
>         for (int i = 0; i < nokPrices.size(); i++) {
>             b.set(0, i, nokPrices.get(i).getAdjustedClose());
>             b.set(1, i, orclPrices.get(i).getAdjustedClose());
>             b.set(2, i, nasdaqPrices.get(i).getAdjustedClose());
>         }
>
> However, I am not sure how to calculate the covariance, i.e. is  
> there  a function for this in a util library, or if now how can I  
> copy the array, take the transpose, subtract the column/row means  
> and multiply the arrays (i.e. the operations necessary to calculate  
> covariance).


Look at org.ojalgo.series.CoordinationSet, and then at  
org.ojalgo.random.SampleSet.

Maybe there could be a util method that takes you straight from a  
CoordinationSet to a covariance matrix, but at this moment there is not.

/Anders


> Thanks,
> Alok
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry(R) Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart  
> your
> developing skills, take BlackBerry mobile applications to market and  
> stay
> ahead of the curve. Join us from November 9 - 12, 2009. Register now!
> http://p.sf.net/sfu/devconference_______________________________________________
> ojAlgo-user mailing list
> ojA...@li...
> https://lists.sourceforge.net/lists/listinfo/ojalgo-user