Application of Time Series to Predict Software Reliability

Abstract

Choosing a software manufacturer is a serious challenge for the customer. Since we cannot apply the test methods of classic production processes in the case of software, we must therefore examine the manufacturer. It is quite clear that a specific manufacturer must have appropriate qualifications. But what if there are several manufacturers with the same certification, and you have to choose between them. In this article, we propose a method that examines the general development activities of manufacturers under appropriate restrictions based on the operational and development errors made, arranging them in time series. The article describes the method, its advantages, limitations, obstacles to its application and an unexpected consequence in education.