Analysis of Conflicting User Requirements in Web Applications Using Graph Transformation

Abstract

User satisfaction is recognized as an important contributor to the success of software applications. It is subjective and influenced by several factors that are linked to the non-functional requirements. Although non-functional requirements provide good criteria for selection of web service in Service-Oriented Architecture, specifying them during the discovery process is a difficult task. Dealing with the non-functional requirements that have an inherent ability to conflict is a complex task and modeling them can facilitate analyzing them. The aim of our research work is to propose a framework to improve design stage during web service selection so as to produce minimal conflicts and dependencies between requirements. We argue that the design structure that best supports and embodies the user constraints is the one that best meets a user's needs. Our framework is proposed after a detailed study of the issues and challenges identified by published literature by researchers working on non-functional requirement To exemplify our work we consider a Remote Patient Monitoring System scenario and propose a complete framework that will take the user requirements as input and analyze how they interact using graph transformation rules. The output matrix of dependencies can help the designer to select the most desirable design solution. We make use of simple natural language processing techniques, graph transformation and the aspect oriented paradigm. By integrating the framework into the design phase, it is possible to limit the impact of changes and also to understand in advance the likely interdependencies. conflicts.

Semi-automated Software Requirements Categorisation using Machine Learning Algorithms

“…Inverse Document Frequency measures the informativeness of word (w). It is calculated by (3) Word2Vec -Word2Vec's objective is to align the vectors of words with similar meanings in the vector space. It searches for mathematical similarities.…”

“…These requirements are ordered as per their significance, such as conflict, penalty, and price [3]. Given the rising demand for software functionality, many software solutions have a wide range of requirements.…”

“…Significant research is being conducted in the field of applying machine learning models and techniques in classifying text [7] from documents and other natural language processing techniques. Based on several studies [3,8,9] and comparative analysis, in this paper, TF-IDF and Word2Vec techniques have been employed for feature selection and several machine learning algorithms, which are Decision tree, Random Forest, Linear regression, Neural network, KNN (K -nearest neighbours), SVM (Support vector machines) for text categorisation.…”