Unified Modeling Language(UML) generation from userrequirements in naturallanguage

Sammanfattning: A significant challenge in software engineering is the process of translating natural language requirements into Unified Modeling Language (UML) diagrams. This thesis delves into the methodologies and approaches involved in the automated derivation of UML diagrams from textual requirements. Furthermore, a software solution has been developed to autonomously generate UML use case diagrams. The primary focus is on two essential tasks: Named Entity Recognition (NER) for identifying relevant entities, and Multiclass Classification for determining another entity in the use case. The study begins with a deep dive into understanding machine learning models and how to evaluate their performance. This is exemplified through illustrative instances showcasing accurate entityidentification and logical UML diagrams. In conclusion, this thesis offers athorough framework for bridging the gap between formal UML diagrams andnatural language requirements, emphasizing the potential of machinelearning approaches for advancing software development and systemdesign