An Efficient Log File Analysis Algorithm Using Binary-based Data Structure

Log files provide valuable insight to previous history of system's usages. Using the information from a log file can help improve future access of a system. However, log files often contain huge amount of data which require significant amount of time to be processed. Even though some popular algorithms already exist to handle this, it is still an open challenge for researchers to further improve them. Hence, in this paper, a novel binary-based approach for frequency mining of a database log file is presented. The approach includes the use of a new algorithms along with its supportive data structures. Construction of the approach began with evaluation of some of the existing methods and identifying their drawbacks. From there the new algorithms were developed and tested. Initial experimentation of the approach reveals a significant improvement in terms of the execution time of the log file's frequency mining calculation.