Abstract
Due to the change of consumption habits of today's consumers, the purchase of products or service is not just to meet the needs of daily life. Thus, how to stand out in a diversified selection of products becomes an important issue. Many scholars explore the usability of products or services to enhance the user experience. Thus the concept of user-centered design arises. In this internet era, web usability has been taken seriously, and thus web mining, web user behavior tracking and other methods are developed to analyze the behavior of web users. To date, scholars have proposed many usability testing methods, but most are artificial methods. In recent years, scholars began to focus on automated usability testing method, because automation can bring in many benefits, such as reducing cost and increasing testing coverage, etc. Although scholars have highly valued the automated testing method, it is rarely used by the organizations due to it needs high technical support. Hence, how to develop an automated testing method without the high cost and high-level support becomes a crucial issue.This study aims to develop an automated usability testing method using existing services and software. First, this study uses automated techniques to record the web user behavior, and to define the user behavior pattern based on fuzzy concept. Then, we do the sequential pattern discovery by web mining technique, and calculate the differences between patterns. Finally, through the sequence alignment, we can find some information about the cognition gap between users and designers. Finally, we construct a shopping website to present a case study using this automated testing method. By conducting simulation and verification of this method, we conclude that this method can enable designers to better construct and improve their website to achieve the purpose of improving web usability.