Master's Thesis Anupama Sajwan

User-Adaptable Rule-Based Natural Language Generation for Regression Testing

Abstract

In today's environment, domain experts are ooded with data, as their job is to analyze data and provide feedback to peers or other people. Their task is too laborious and time-consuming, therefore they need a software which can help them with their work by automating it. The focus of this thesis is to carry out a research on current state of art in eld of Natural Language Generation (NLG). In addition, implement a pilot project which can automate the process of data analysis and expresses the result in form of textual reports. Case under investigation is related to the nancial domain, hence the developed tool's logic should be transparent to business users; it should not be a black box. To increase the eectiveness of the tool, it should be user-adaptable. In the end, the developed tool is evaluated to check its eectiveness and user acceptance.