Looking for the latest version? Download scomposer3.zip (379.6 kB)
Home
Name Modified Size Downloads / Week Status
Totals: 2 Items   383.4 kB 2
scomposer3.zip 2012-12-03 379.6 kB 11 weekly downloads
README.txt 2012-11-26 3.9 kB 11 weekly downloads
How to run this program? You should have the JAVA for developers installed. Extract it into a desired directory (from which you can run JAVA programs). Compile it with "java test.java" Then start it with "java test" The result will be stored in a "out.txt" file, which is to be found in the same directory as the program files. What is this program? This program is meant to search the Internet and suggest sentences that might follow any inputted sentence. It was meant to help composing the writings based on Internet search. How was it made? First, I made a chatterbot to understand the concept of "predicting" the next sentence (worked as good as the Cleverbot), then I made this program that is aided by an user to compose the new writings. Bugs? Yes, there are, and they are not mine but came from Jsoup classes. Sometimes the Internet connection simply takes too much time to get the data. This has to be fixed. Future releases? None will be made, I made this program just for fun, so you may use this program however you like... make something new from it etc. There are several parameters that must be inputted: Prefix: This is something that is always included in the internet search, you may use internet search booleans and so on. This "focuses" the search. Sentence number: The number of sentence choices that program should offer. Input sentence: This is the sentence that program considers while searching for the next most suitable sentence. The result is stored inside an output.txt file. The sample run: The prefix was set to be "recipes" The first sentence (inputted by me was) "Boil two eggs". Running the software and choosing the most appropriate sentences (in my unprofessional cooking opinion) gave the following result: Boil two eggs. Remove from hot water, cool, peel and chop. Place the chopped eggs in a bowl, and stir in the mayonnaise, mustard and green onion. Season with salt, pepper and paprika. Stir to blend, then cover and simmer over low heat for at least 2 hours, stirring occasionally. After 2 hours, taste, and adjust salt, pepper, and chili powder if necessary. Just before serving, add lime juice and cilantro. To check the validity of the program, I have made the recipe. Since I didn't have any green onion, I used cardamom, garlic powder and a dried/smashed forest mushroom. Everything else was done as written. It looked like a shit served on a bun, the taste was too sour for me, so either I put too much mustard or lemon juice. The greacy aftertaste from the mayonnaise was too heavy and made me want to puke... I also had to wake up at 4AM since I didn't feel good having an aftertaste in my stomach (almost threw up) but carbonated water fixed my stomach. I call this recipe, "The taste of the Internet"... the love for AI is hard. Photo as a proof: http://damir-olejar.deviantart.com/#/d5m9wiz For the second run, I have chosen something that is more close to my field of interest, and the result was the following: Input sentence was "Learning algorithm" The prefix was "transductive learning" Learning algorithm. Transduction, or transductive inference, tries to predict new outputs on specific and fixed (test) cases from observed, specific (training) cases. The simplest realization for transductive inference is the method of k-nearest neighbors. Learning to learn learns its own inductive bias based on previous experience. Because training sets are finite and the future is uncertain, learning theory usually does not yield guarantees of the performance of algorithms. Instead, probabilistic bounds on the performance are quite common. In addition to performance bounds, computational learning theorists study the time complexity and feasibility of learning. In learning theory, a computation is considered feasible if it can be done in polynomial time. ....AND SO ON.... Enjoy the program !
Source: README.txt, updated 2012-11-26