Natural Language Processing Using Database Context

Abstract

The usage of natural language in the industry has become more prevalent in recent years. Nowadays it is much easier to operate with complex infrastructures using natural language. This subfield of artificial intelligence is becoming more widespread every year. Feature semantic parsing represents one of the tasks of converting natural language utterances into structured logical parts that can be used as queries to generate responses. This paper introduces an algorithm that transforms natural sentences to obtain structural results. Such a functionality is effective for Question and Answering (Q&A) and allows for spoken language understanding. Obtaining the input as natural language sentence allows many people, even those without technical skills to access information effectively. The generated queries are being executed on specific tables or databases and then the information is retrieved in a comprehensible way. The process of generating queries from natural language sentence implies different operational steps, such as recognition of different types of words, synonym detection, feature grammar parsing, etc.