This is our code for NDBCCUP2015 on knowledge Extraction from the web - Semantic Relevancy Calculation of Short Texts, and we got Third-Class Award (4/92). The mean squared error of our solution is 0.772697, and the best result is 0.737663.
Team member: Jiaming Xu, Jun Tian, Shiyu Zhou.
Adviser: Bo Xu, Guanhua Tian, Fangyuan Wang.
Please feel free to send me emails (jacoxu@msn.com) if you have any problems.
/2015Sougou_XTZ_results: transform to submission format.
/Data: We do not share the data, please find it on the official website.
/feaSet: all features are saved in this folder.
/javaCode: java code for text process.
/library: all external sources.
/matlabCode: matlab code for classification.
/20150501-提交版-短文本语义相关度计算说明文档.pdf
