huilyu2 / word-sense-disambiguation

CIS520 Spring 2018 Final Group Project

Data Preprocess

Due to the large volumn of the generated dataset, we didn't upload the dataset to this repository. The following scripts generate the dataset:

parser.py generates the dataset for multi-class classification, from SemCor corpus.

split_dataset.py splits the dataset produced by parser into training set and test set, in 70/30 ratio. The split is stratified to account for class imbalance.

Learning Algorithms

bayes.py: Naive Bayes

logistic_regression.ipynb: Logistic Regression

svm.ipynb: Support Vector Machine

waveNet_keras/train.py : Neural Network