Classifying voice-based customer query using machine learning technique

Title

Classifying voice-based customer query using machine learning technique

Creator

Kapse M.; Elangovan N.; Varghese A.; Joy A.; Tom C.

Description

Timely attention to issues raised by customers is critical. It is imperative that the average handling time is lesser, which in turn contributes to productivity. It was found from the data from the banking industry in the US that, on average, a customer service call last for seven minutes. The first two minutes are for the call to get redirected to the respective team. This study investigates a method using machine learning to classify and redirect the customers into the respective department directly based on their initial voice response or voice message. It will substantially reduce the service time. CRISP-DM methodology is being used to design the process of the study. The most frequently occurring issues and the department to which they are associated are created through machine learning from the dataset that contained product reviews and metadata of different issues. The programming languages that are used in this study are Python, HTML and Java. An interface is created by using HTML, which makes it quite user-friendly. The study tests the effectiveness of converting voice to text and interprets which department the call should be transferred to address the issue. A support vector machine and a logistic regression model were used for the prediction, and it was found that the models provided an accuracy of 83 and 84 percent, respectively. The study proves that using ML and voice recognition reduces the average handling time. 2021 Ecological Society of India. All rights reserved.

Source

Indian Journal of Ecology, Vol-48, pp. 69-72.

Date

2021-01-01

Publisher

Ecological Society of India

Subject

Feature extraction; Stemming; Tokenization; Vectorization; Voice recognition

Coverage

Kapse M., School of Business and Management, Christ (Deemed to be University, Bengaluru, 560 029, India; Elangovan N., School of Business and Management, Christ (Deemed to be University, Bengaluru, 560 029, India; Varghese A., School of Business and Management, Christ (Deemed to be University, Bengaluru, 560 029, India; Joy A., St Joseph's College of Arts and Science (Autonomous), Bengaluru, 660 027, India; Tom C., St Joseph's College of Arts and Science (Autonomous), Bengaluru, 660 027, India

Rights

Restricted Access

Relation

ISSN: 3045250

Format

Online

Language

English

Type

Article

Identifier

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85113799410&partnerID=40&md5=44e4f957c86c2241c77b893bc18806d9