Software Interactive Optimization Based On User’s Behavior Sequence

With the rapid development of information technology, the competition between functionally similar software become intense. In order to stand out from competitors, the software product should improve software performance and the interfacial interaction on the basis of the basic functions.At the time of the software design is nearing completion or has been completed, this paper collect user’s logging data for research, hoping to find out the critical path of the user’s behavior. Then analysis user’s behavior and motivation target based on the order of operating software function. Adjust the software interface with user interface design principles and guidelines, in order to get the appropriate interface effects. The main work is as follows:First introduce research background, research content, course-related research development and future prospects, and introduce the main work, user behavior analysis and sequential pattern mining theory.Secondly, user’s behavior log data is used in this paper to do preprocessing and statistical analysis, including delete outliers, invalid clicks, sort specific attributes, split sessions, calculate and show user behavior indicators, such as the number of functions operate, function interval and session sequence length.The main content of this paper include mining intervals action sequences, non-interval sequential patterns two parts. In the process of mining interval function sequences, this paper first produce session matrix, use cosine function to generate the associated set of clusters. Input the correlation function into GSP sequence mining algorithm to generate a sequence pattern results. In the process of mining the specified length sequence patterns of non-intervals sequence patterns, compute support and sort to find the highest support items, analyze the general trend of non-intervals sequence patterns.Finally, the sequence of results do visualization, and combined statistical results, intervals sequential pattern and non-interval sequential pattern, analyze the causes of user behavior and identify possible defects in the software architecture. Summarize the content of the work and the direction of future work.