Software Complexity Prediction by Using Basic Attributes

Rasha Gaffer. M. Helali

Software complexity, LOC, McCabe, halstead, branch count.

Software complexity is one of the important quality attribute that affect the success of software. Predicting such attribute is a difficult task for software engineers. Current used measures for computing complexity are not sufficient. Data mining can be applied to software data to explore useful interesting patterns. In this paper we present a simple data mining based prediction model to predict software complexity based on some basic attributes. The article starts by considering the correlation between different features that describes software code structure then selecting some of these features to be used for complexity prediction. Results reveal the ability to use branching count feature as strong predictor of complexity.

10.22161/ijaers/3.11.21