| Summary: | 碩士 === 真理大學 === 管理科學研究所 === 93 === Search engine have turned into major tools in helping users’ searching for their suitable information from the Internet. Search engines are designed to efficiently filtering out the unsuitable information, but we usually still confront too much unsuitable information when using the search engine. In other words, the search engine users still need spending lots of time to find their favorable web pages. In 2002, Picarougne et al. proposed a GA-based algorithm named Geniminer, and hoped that Geniminer will effectively find users’ preferable web pages. We supposed that Geniminer cannot bring the genetic algorithm’s creativities into full play, so we propose an alternative algorithm named Gaminer to increase the creativities of the Geniminer. To verify the performance of Gaminer, we finish both Gaminer-based system and Geniminer-based system and compare Gaminer-based system with Geniminer-based system on precision ratios, recall ratios, satisfactions and creativities. The experimental results of five cases show that the proposed algorithms outperformed Geniminer on precision ratios, recall ratios and satisfactions but no significant difference on creativities. According to the evidence, we conclude that the proposed algorithm is feasible for helping internet users searching for their satisfying web pages, but how to increase creativities of the Gaminer is still remained to be solved.
|
|---|
