2010 年 12 巻 4 号 p. 417-426
Our goal is to point out problems in web pages with low usability in order to improve the web usability. We investigate the relation between user interaction behaviors in web-viewing and evaluation results of web-usability by subjects, and extract discriminative patterns for user interaction behaviors in viewing of web pages with low usability by using the PrefixSpan based subsequence boosting (Pboost). In the experiments using 63,000 records of the user interaction behaviors in viewing of 263 web pages which include 20 pages with low usability, we show that 76 discriminative patterns of user interaction behaviors are effectively extracted.