An improved software fault prediction schemes for object oriented system

Abstract

Object oriented software provide ease environment for implementing and accomplishing real world task easily by providing more objective ideas. OOPS can ensure the reliable and qualified software development in which more improved and competent software can be implemented. However, faults and bugs arise in the object oriented program which is lead to inefficient software implementation. The main goal of this research work is to implement the software research methodologies which can identify the faults present in the software more efficiently, so that efficient software construction can be assured. In the first research work, the Fault Tree Analysis (FTA) method is introduced which focus on predicting the software faults in the accurate way. The quantitative approach utilizes probability information correlated with every condition to estimate the likelihood of the happening of that particular state. By considering comments and identifiers the fault is predicted by the use of conceptual coupling and cohesion methods. It was created by the developer for future references. If the developer does not aware of comments it will produce lack of measurement. In this method the fault predicted after the coding phase. In software fault tree analysis, the fault is predicted in requirements, design and implementation phases. But it considers only approximated failure cases during the fault prediction. In the second research work it determines the new malicious code identification algorithm. It will minimize the unknown code in the developed program, enhances the quality of software, reusability and easy debugging. In this research, the information is drawn-out into OpCode n-gram representation that will assist to detect the unknown code. newline newline newline