SISRR: Semantically Inclined Strategic Learning Model for Software Requirement Recommendation Using Artificial Intelligence

The following paper proposes the SISRR framework for knowledge centric, semantically inclined, framework for software requirement recommendations. This framework is intelligent driven by integrating semantically inclined techniques with knowledge model. The SISRR model is a metadata driven framework, where the initial requirements are preprocessed, and metadata is generated to intensify the auxiliary knowledge. This is handled by deep learning GRU classifier. Subsequently ontologies are generated from the SVN repository, which is applied along with Bag of Words to enrich the ontologies. The semantic similarity is computed using CoSimRank and Tang Index. Shannons entropy is also computed to measure the information content. The context map is generated using the DBSCAN algorithm in the model. Thus, the best precision, recall and lowest FDR has been achieved by the SISRR model compared to other existing models.