The Inverse Edit Term Frequency for Informal Word Conversion Using Soundex for Analysis of Customer’s Reviews

Abstract

Background: E-commerce/ M-commerce has emerged as a new way of doing businesses in the present world which requires an understanding of the customer’s needs with the utmost precision and appropriateness. With the advent of technology, mobile devices have become vital tools in today’s world. In fact, smart phones have changed the way of communication. The user can access any information on a single click. Text messages have become the basic channel of communication for interaction. The use of informal text messages by the customers has created a challenge for the business segments in terms of creating a gap pertaining to the actual requirement of the customers due to the inappropriate representation of it's need by using short message service in an informal manner. Objective: The informally written text messages have become a center of attraction for researchers to analyze and normalize such textual data. In this paper, the SMS data have been analyzed for information retrieval using Soundex Phonetic algorithm and its variations. Methods: Two datasets have been considered, SMS- based FAQ of FIRE 2012 and self-generated survey dataset have been tested for evaluating the performance of the proposed Soundex Phonetic algorithm. Results: It has been observed that by applying Soundex with Inverse Edit Term Frequency, the lexical similarity between the SMS word and Natural language text has been significantly improved. The results have been shown to prove the work. Conclusion: Soundex with Inverse Edit Term Frequency Distribution algorithm is best suited among the various variations of Soundex. This algorithm normalizes the informally written text and gets the exact match from the bag of words.