Latent Relevance Tools (LIBLATREL)

@brief:	to train latent relevance model, and test given test_file
@wiki:	http://wiki.babel.baidu.com/twiki/bin/view/Com/Ecom/LatentRelevance
@email:	lichangcheng@baidu.com

Training
Usage: ./train [options] training_file [model_file]
options:
	-f mode: 1 with factorization, 0 no factorizatoin (default 1)
	-k factorization dimension (default 3)
	-c regularization coefficient (default 0.1)
training_file format: 
	label \t x1 \t x2 (example: 1 \t 0.1 0.2 \t 0.3 0.1)

Notice: use "-f 1" for dense vector model (word2vec, etc.).

Testing
Usage: ./test test_file model_file
test_file format: label \t x1 \t x2 (example: 1 \t 0.1 0.2 \t 0.3 0.1)

Predict results are saved in ${test_file}.res. You can run "./auc -r ${test_file}.res" to get the AUC of the test_file.

For any questions and comments, please send your email to lichangcheng@baidu.com