Shallow parsing natural language processing implementation for intelligent automatic customer service system

Abstract

This paper introduces implementation of Shallow-Parsing methods in Question Answering System of Natural Language Processing for Indonesian Automatic Customer Service. The main idea of the approach using Shallow-Parsing is indexing the answer. In this paper we present some steps of simple Information Extraction (IE) using Shallow-Parsing such as Part-of-Speech Tagging, IOB Tagging, Text Chunking, Predictive Annotation, Relation Finding, and Answer Retrieval. The main purpose of the task is to identify key information, key phrase and question phrase that contained in each question or answer. With that, information system can index the given question and retrieve the relevant answer to customer. The experiments of Automatic Customer Service reported in this paper show competitive results, given 100 questions; system can respond 89 questions with relevant answer correctly. This experiment shows that the accuracy of the Automatic Customer Service system is 89% of 100 given questions.