valid-method-ids-cnn

First, preprocess data for deep learning.

1. append 0 to all vectors, make all vectors hold the same size.

Input: inputData/unsupervised-learning/ Output: outputData/WithoutNormalization/

2. Normalize the values in the data vectors of results of the first step.

Input: outputData/WithoutNormalization/ Output: outputData/Normalization/

3. Standardize the values in the data vectors of results of the first step.

Input: inputData/unsupervised-learning/ Output: outputData/Standardization/

Second, extract(encode) features of data vectors by deep learning (CNN).

###1. extract(encode) features of method body. Input: src/main/resources/WithoutNormalization/, src/main/resources/src/main/resources/WithoutNormalization/, src/main/resources/Standardization/ Output: outputData/CNN/

2. extract(encode) features of method name.

Input: src/main/resources/WithoutNormalization/, src/main/resources/src/main/resources/WithoutNormalization/, src/main/resources/Standardization/ Output: outputData/CNN/

The results of CNN will be used in clustering.