ID3 Algorithm

implemented in Java

---

EVANGELOS PIPILIKAS

File compilation

1. Open the command line
2. Change the directory to the one that contains the project
3. Run the command: javac ID3.java

Running the algorithm

>!< Please pay attention to the order that parameters should be given. >!<

• To execute the algorithm, make sure you have put the train and data files at the same folder with algorithm ( or make sure you put the right directory).
• The algorithm will expect the above parameters with the exact same order that are described:
  • featureFileName: the file name that contains the vocabulary (.vocab file)
  • trainDataFileName: the file name of the training examples (.feat file)
  • testDataFileName: the file name of the test examples (.feat file)
  • pruningParameter: the parameter we use to prune the tree (double and should be between 0.0 and 1.0, we recommend > 0.80 and <0.97)
  • nParameter: the first n most common words
• Here is an example of input: java ID3 imdb.vocab trainlabeledBow.feat testlabeledBow.feat 0.90 100

>!< In case there is an OutOfMemoryError, please use less words in the .vocab file. >!<

Source of data in test_examples folder: https://ai.stanford.edu/~amaas/data/sentiment/