Soft computing detection method for input data error problem in computerized information system: A framework

Abstract

A detection method for input data error in computerized information system is a critical study because wrong inputs will cause incorrect results. Although some errors can be controlled through the interface input design, yet there is no guarantee that the potential error of inputted data is detected. Therefore, there is a need to have a tool which embeds intelligent data analysis to check the potential errors of the inputted data. By having this tool it will provides an early warning of potential error by detecting any abnormal behavior of the inputted data. This paper proposes a framework that used supervised learning neural network that functions as an intelligent data analysis to analyze the behavior of inputted data to detect potential error for computerized information system. To apply the framework, the research uses student information system that store students’ marks and calculate their cumulative point average (CPA). In student information system there is a potential of CPA is incorrectly calculated due to mistaken inputted marks and consequently will effect the achievement of the students and causes doubt to the university’s assessment system. Therefore, this paper presents details on how the framework could overcome this problem by detecting abnormal pattern of entered marks.