Searching Experience Sharing Based on Ant Colony Model

• Main Authors: Chung-Jen Nien, 粘仲仁
• Other Authors: Jung-Sing Jwo
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/89227330180021029098

碩士 === 東海大學 === 資訊工程與科學系 === 97 === Search engine is an important tool to the current web users. It helps users finding useful contents from a huge repository. However, search experience is difficult to be generated and reused. In the thesis, a concept of search-trails and ST-Map are proposed. Instead of just considering the search targets, search-trail focuses also on the search process. The trails built in search process are very similar to the trails created in an ant colony. Therefore, based on ant colony model, ST-Map is designed to help web users sharing search experience with each other. In order to see how ST-Map influences the search behaviors, a set of simulations is constructed. In the simulation scenario, expected search-trails are assumed to be known. In these simulations, even for the extreme case, with 2% experienced, 8% standard, and 90% inexperienced users, they collaboratively generate the desired search-trails with 96.29% similarity of the expected ones in two months period. For the other two cases, the search-trails with 100% similarity of the expected ones can be generated in less than one month period. These simulations show that ST-Map can really help users sharing and reusing the search experiences. Furthermore, these experiences are polished day by day just like the ants can find their trails to the target food sources. Keywords: search experience sharing, ant colony, pheromone, search-trails, ST-Map.