Machine Learning Project 2 Luke Hanks Trevor La Pay Kyle Jurney
The following files need to be downloaded from kaggle first: testing.csv training.csv
LOADING DATA: The code tries to load the basic data (ALL_CLASSES, ALL_WORDS, ALL_CLASS_EXAMPLES, TEST_EXAMPLES) from a pickle file called "ConstructionZone2Vars.pkl" which takes less than a second. If that doesn't work, then it will try to load the data from the original files (found on Kaggle) which takes about 30 minutes. If the data is loaded from the original files, the script will create the "ConstructionZone2Vars.pkl" pickle file for next time.
To run Naive Bayes classification, Logistic Regression, or any combination of those and the sub-problems we solved in this project, find the boolean flags at page top of the file ML_2018_NaiveBayes_LogisticRegression.py and modify what you want to run to True:
DO_NAIVE_BAYES = False DO_NAIVE_BAYES_BETA_SEARCHING = False DO_LOGISTIC_REGRESSION = True DO_LOGISTIC_REGRESSION_NUM_INTERS_SEARCH = False DO_LOGISTIC_REGRESSION_LEARN_RATE_SEARCH = False DO_LOGISTIC_REGRESSION_PENALTY_SEARCH = False DO_LOGISTIC_REGRESSION_VALIDATE = True DO_LOGISTIC_REGRESSION_TEST = True