Scientific Literature Retrieval Model Based on Weighted Term Frequency

摘要

Science and technology literature retrieval system is commonly used by researchers as document retrieval tools. The classic information retrieval models such as Boolean Model [1], Vector Space Model and Probabilistic model neglect the position of query words in every paper. We propose a weighted term frequency model (WTFM) based on term frequency, and through a simulated annealing algorithm to learn the weighted factor. The results of experiments show that our weighted factors model gets better performance than ordinary models.