Analysis of BM25 Parameters in Queries Related to Food and Drug Companies
50 likes | 174 Vues
This project explores the application of the Okapi BM25 algorithm in processing queries about food companies and bankrupted drug firms in China. By analyzing the effect of parameters k1 and b in BM25, the research evaluates their impact on ranking document relevance. The results include document ID lists for specific queries such as George Bush and healthcare reform policies. This analysis sheds light on the effectiveness of BM25 in retrieving relevant information regarding significant corporate events and political topics.
Analysis of BM25 Parameters in Queries Related to Food and Drug Companies
E N D
Presentation Transcript
COMP6791 Project2 Yuan Tao
k1 and b in BM25 Query: food company china
Sample queries • Drug company bankruptcies Document ID list (In total: 2): 3091 (14.2894) 2127 (11.8888) • George Bush Document ID list (In total: 13): 20891 (15.9528) 8593 (13.1173) 4008 (12.3950) 16780 (11.4868) 20719 (10.4883) 3560 (10.2646) 7525 ( 9.1157) 20860 ( 8.9045) 8500 ( 8.8002) 2796 ( 7.7240) 965 ( 6.0912) 5405 ( 6.0723) 854 ( 6.0556) • Democrats' welfare and healthcare reform policies • Democrats welfare (3) • Healthcare (30) • reform policies (109) • Democrats welfare policies (1) • Democrats welfare reform (1)