Research on Scoring Mechanism Based on BM25F Model
Guangzhong Lan, Ge Yunsheng, Jie Kong · 2019
The BM25F correlation sorting algorithm is mainly implemented on the basis of the BM25 algorithm. However, it failed to take into account the impact of multiple other impact factors on the ranking of search results. In order to effectively solve this problem, According to the comprehensive influence factors of the journals, Reader's clicks, downloads, and citations for articles, Article published by the time period, The influence of long vocabulary and short vocabulary on the ranking of retrieval results is also considered. And set weight parameters for it, Then merge with the BM25F search result sorting algorithm formula. So as to improve and optimize the score of retrieval results. It is proved by the experimental results. The improved BM25F sorting algorithm has the retrieval effect. It can be more consistent with the needs of the vast number of users. It enables articles with relevant frontiers to get higher scores.