An efficient logistic regression and ant colony optimization-based object-oriented quality prediction

In this paper an efficient logistic regression and ant colony optimization-based object-oriented quality prediction has been presented. The dataset has been considered based on the object-oriented codes. The code considered here are completely equipped with object-oriented features for the complete analysis. The major parameters considered for the experimentation are class, object, inheritance and dynamic behavior. Class and objects have been considered for the class labels and memory requirements checking with the proper stack call and constructor invocation. Inheritance has been considered for the code reusability testing. Dynamic behavior has been tested for the runtime allocation. Then clustering algorithm is used for the parameter preprocessing and grouping based on the parameters. For data filtration chi-square testing has been applied. Then logistic regression and ant colony optimization (LR-ACO) have been considered for the final classification. Then F-Measure, Odd Ratio and Power have been considered for the analysis of the classification based on LR-ACO. The result after LR-ACO shows better accuracy as comparison to the previous methods.