Quality Assessment for Requirements based on Natural Language Processing

Recently, many approaches for automatic information extraction from technical specifications in the area of electronic design automation have been proposed. For this purpose, techniques from natural language processing are used. In order to lower the bars for designers and customers, some approaches do not intend to restrict the natural language that is used to describe the specifications. However, ambiguity and vagueness in the language often cause wrong or bad results obtained from the algorithms. This work describes preliminary ideas for automatic approaches that assist in writing the specification and aim at increasing the quality of the text. Besides improved comprehensibility, better written specifications will also enhance the quality of the automatic extraction approaches in subsequent steps of the design flow.