/SBMLT

Primary LanguagePython

SBMLT

##Folders

  • "cooking_code" folder : All of the code (and the input) goes here. We use scikit-learn's implementation of our learning algorithms.

    • cooking_util contains reusable code and is meant to be imported as a module. It includes:
      • a function that loads the data
      • a function that cleans the data using Elad's cleaning method
      • a class that makes the data into numpy arrays
      • a function that creates a submission file, using a prediction function
    • log_reg.py uses scikit-learn's logistic regression with Elad's cleaning method
    • random_forest.py uses scikit-learn's random forest with Elad's cleaning method
    • log_reg_bag_of_words.py uses scikit-learn's logistic regression with the bag of words cleaning method
    • random_forest_bag_of_words.py uses scikit-learn's random_forest with the bag of words cleaning method
  • "cooking_output” folder : Submission files generated by the algorithms

    • kaggle_scores.txt stores the csv files' accuracies on submission to Kaggle.