ListMLE.rar_learning to rank_listmle

Author: Ngo Xuan Bach ([email protected]) Date: April, 2012. Version: v1.0 ====================================== 1. OVERVIEW ListMLE is a tool for learning-to-rank. It is implemented based on the ListMLE algorithm: Source paper: Listwise approach to learning to rank: theory and algorithm (ICML 2008) =============================================================================================================================================== 2. HOW TO USE 2.1. Binary a) learn Usage: learn data_file tolerance_rate learning_rate model_file b) multiLearn When the training file is too big, we need to divide the file into some smaller files and put them into a directory. In this case, we use multiLearn.exe instead of using learn.exe. Usage: multiLearn dir_path tolerance_rate learning_rate model_file dir_path: the directory contains all training files c) test Usage: test model_file data_file output_file 2.2. Build Please make a C++ project in MS Visual Studio and compile. ================================================================== 3. FILE FORMAT (TRAIN/TEST) <line> .=. <target> qid:<qid> <feature>:<value> <feature>:<value> ... <feature>:<value> <target> .=. <float> <qid> .=. <positive integer> <feature> .=. <positive integer> <value> .=. <float> Here is an example: 3 qid:1 1:1 2:1 4:0.2 2 qid:1 3:1 4:0.1 5:1 1 qid:1 2:1 4:0.4 1 qid:1 3:1 4:0.3 1 qid:2 3:1 4:0.2 2 qid:2 1:1 2:1 3:1 4:0.4 1 qid:2 3:1 4:0.1 1 qid:2 3:1 4:0.2 5:1 2 qid:3 3:1 4:0.1 5:1 3 qid:3 1:1 2:1 4:0.3 4 qid:3 1:1 4:0.4 5:1 1 qid:3 2:1 3:1 4:0.5 For more information, please contact me at: [email protected]