Abstract: Web usage mining plays an important role in the personalization of Web services, adaptation of Web sites and the improvement of Web server performance. It applies data mining techniques to discover Web access patterns from Web usage data. In order to discover access patterns, Web usage data is reconstructed into sessions with or without user identification. However, not all Web server logs contain complete information for constructing user sessions. It has been proposed three extensions to heuristics in order to improve their performance. This study describes improved statistical-based time oriented heuristics for the reconstruction of user sessions from a server log. The improvement caused by using the heuristics can thus be ascertained by multiple trials with same data set.