20news

text classification on 20 newsgroup, using model of perceptron, naive bayes, knn and svm

1.unzip "data.zip"

The final root directory should contain files: data_2, data_3, data_10, pre_processing, main.py, my_bayes.py, my_knn.py, my_perceptron.py, my_svm.py

2.run the command

run python main.py --class_num 10 --model bayes or simply python main.py (the default number of classes is 10, the default model is bayes) in the command line