摘要: University of Chinese Academy of Sciences (UCAS) participated in the first task, namely temporally-anchored ad hoc retrieval in Microblog track, aiming to efficiently and effectively retrieve tweets. Based on the conventional appl... 展开 University of Chinese Academy of Sciences (UCAS) participated in the first task, namely temporally-anchored ad hoc retrieval in Microblog track, aiming to efficiently and effectively retrieve tweets. Based on the conventional application of learning to rank, we incorporated a machine learning approach, such as logistic regression for selecting high-quality training data for improving the effectiveness. Except for the tweets' content features, we also used the features of the web information, external evidence, which is related with the URLS to improve the effectiveness. 收起