manishmazumder / craftml Goto Github PK

This implementation is meant to work with ‘Delicious’ dataset. The data set can be downloaded from (“http://manikvarma.org/downloads/XC/XMLRepository.html”) and can the code and the dataset should be kept in the same folder before the code can be executed. 

If you want to execute on some other data set then change lines 329 and 560 and then make sure that training data is stored in two matrices ‘train_data_X’ and ‘train_data_Y’. Similarly the validation data must be stored in ‘validation_data_X’  and ‘validation_data_Y’ and the test data should be stored in ‘test_data_X’ and ‘test_data_Y’. 

The file ‘CRAFTML_delicious.py’ can be opened using any python editor like spyder and then can be run. 

This code shows the training time and the accuracies.