Search Pattern Analysis and Evaluation for a Subject-oriented Search Application

• Main Authors: HUANG,YEN-CHUN, 黃彥鈞
• Other Authors: Wu,I-Chin
• Format: Others
• Language: en_US
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/96274464605015223952

碩士 === 輔仁大學 === 資訊管理學系碩士班 === 104 === This paper addresses ways of presenting users with lists of results from various websites and in different media types by integrating the search function and information visualization (IV) tools. Based on our previous research with the WikiMap# application, although the topic map tool improves user searches, the interface is too complicated for users to operate easily. In response, we refined the topic map tool based on program theory (PT) and have proposed a new interface, W WikiMap○P. The major difference between Wikimap# and WikiMap○P is the refined topic map tool, which now integrates the semantic path suggestion. Our research involved using extended evaluation measures, zero-order state transition (ZOST), and maximal repeat patterns (MRPs) to observe and analyze users’ search move behaviors with the interface, as well as the relationship between those moves and task accomplishment. Results of evaluation show that users generally preferred the refined topic map tool to search for answers, with which they often used the strategy of selecting the topic map tool as the primary tool and the search tool as an auxiliary one. Users also tended to achieve the best task performance while using the proposed WikiMap○P interface. In all, our research has revealed that users’ search patterns can be identified in different tasks, for which we have proposed search strategies based on the results of search move analysis.