Selecting Software Architectural Styles from Requirement Documents Using NLP

Abstract

With the growth in the size and complexity of softwares, the design phase of software systems has recently drawn more attention. Software architecture is a fundamental idea in this phase and has a significant impact on the software development life cycle, with the degree to which it is utilized frequently determining the success of a software project. Software architecture prediction is the crucial step before the implementation phase. To solve the inherent issues with current methods that have been reported in the literature, this research makes a novel method based on quality attributes to assess software architecture design. The PURE dataset has been used to extract the quality parameters from Software Requirement Specification (SRS) document. Natural Language Processing (NLP) and Machine Learning (ML) techniques have been used to bring automation in architecture styles prediction with efficiency and accuracy. The Automated Architecture Style Prediction (AASP) system has minimized the architect role, and 93.75 accuracy has been achieved through Artificial Neural Network Algorithm. Architecture Style designing can be automated in future and can consider more architecture styles.