Preprocessing.py : initialize and processes the input data , if you need to use the word2vec model

construct.py: different classifiers are defined here

model.py: main function entry implementing different kinds of active learning approaches

To run these algorithms

python model.py 

Parameters:

• initial_batch -- set the initial labeled data size

• step -- set the batch size

• model -- set the classifier, e.g.: 'linear', 'logistic', 'bayes'

• methods -- set the active sampling approaches, e.g.: 'active', 'margin', 'center' , 'random', 'entropy' ,'all' means running all these five algorithms

• input_choice -- set the preprocessing methods, e.g.: 'tf-idf', 'word2vec'

• plot -- if true, then five graph of accuracy per batch are plotted, otherwise, accuracies are saved in a txt file

items.json: taobao crawled review

yanjing.anonymous.replace: queried data

Finally, enjoy ur winter vacation www