Designing, developing, and maintaining are the key phases of a software life cycle, and the most essential property of a software product is quality. The quality of a software product is dependent on various factors, e.g., reliability, security, and efficiency. But the most important aspect of a software quality is its proper function as described by functional requirements of the software. Errors often occur that are the mistakes which hamper the correct functionality of the software. Thus, to deliver high-quality software, errors must not occur, and if they do then these must be removed. In this chapter, we suggest that the process of identifying and removing errors can be optimized if prior information about the module’s possible errors is known. Error proneness prediction can be modeled using classification and prediction techniques. In this context, artificial neural network is a classification model which can be used to predict error proneness. However, the neural network with gradient descent algorithm, e.g., backpropagation algorithm, has the inherent issue of getting stuck into local minima while training. To solve this issue, evolutionary algorithms such as genetic algorithm and bird mating algorithm focus on training of artificial neural. When the prediction model is formalized, receiver operating characteristic curve and accuracy curve are used to analyze the performance of the model. In this chapter, we present an error proneness approach using bird mating algorithm.
CITATION STYLE
Pal, A., Jain, H., & Kumar, M. (2017). Optimizing Software Error Proneness Prediction Using Bird Mating Algorithm (pp. 257–287). https://doi.org/10.1007/978-3-319-54325-3_11
Mendeley helps you to discover research relevant for your work.