Extraction of key attributes from natural language requirements specification text

Abstract

The automation of extracting the information from natural language text has become an area of growing interest in the current years. The requirements can be analyzed through extraction process from natural language text which has its own limitations. Software Requirements Specification (SRS) gathers all the requirements that are required for the user. This proposed concept presents an idea to identify the schema for the tables and relationship corresponding to all the tables extracted from the natural language requirements specification. The work starts with identifying the table schema and their properties. Then the Primary Key (PK) attribute is identified based on adjectives, prioritizing the preference of the attributes, hand crafted rules and machine learning system which is trained from statistical data. Next the relationship is identified between the tables using the extracted list of attributes with PK to identify the Foreign Key (FK). The FK attributes of a table is identified by the highly referenced PK attribute and also based on the implicit relationship among the tables. Furthermore, the results presented exhibit the relationship among the tables and the proposed approach is validated using real time data.