A Framework for Selection of Machine Learning Algorithms Based on Performance Metrices and Akaike Information Criteria in Healthcare, Telecommunication, and Marketing Sector

A. K. Hamisu,K. Jasleen

DOI: https://doi.org/10.1201/9781003226147-6

2022-01-12

Abstract:Machine learning (ML) has been a prevailing method in the knowledge discovery process. Recently there has been attention on applying ML algorithms for tasks such as classification. This research aims to propose a framework for the selection of ML algorithms based on dataset attributes, performance parameters, and Akaike information criteria (AIC) score. A total of eight datasets were selected from the healthcare, telecommunication, and marketing sectors. For this experimentation, 13 different ML algorithms were selected and divided as eager learners, lazy learners, and hybrid learners. Evaluation of ML algorithms was carried out on different performance metrics and AIC score. On the basis of accuracy performance metric, results revealed that eager learners performed well for the healthcare, telecommunication, and marketing sector dataset. Average accuracy reported by eager learners in healthcare, telecommunication, and marketing was 90%, 90%, and 94%, respectively. With a receiver operating score of 1.0, the top performing algorithm for telecommunication and marketing datasets was decision tree (DT) whereas support vector machine (SVM) was the best for healthcare datasets. On basis of the AIC, in the marketing dataset, the lowest score was reported by SVM and in the telecommunication as well as healthcare datasets, whereas the lowest score was reported by k-Nearest Neighbor (KNN). This chapter focuses on applications of Machine learning (ML) methodologies in three separate sub-domains: healthcare, marketing, and telecommunications. In the case of the healthcare dataset, support vector machine (SVM) is proven to be the best ML algorithm, whereas in the case of the telecommunication and marketing dataset, DT comes out as the top performing one. DT and SVM are the most suitable algorithms for the telecommunication as well as healthcare sectors. On the basis of AIC score, lazy learner category of ML algorithms is the best option for the telecommunication and healthcare sectors. ML has been a prevailing method in the knowledge discovery process. A total of eight datasets were selected from the healthcare, telecommunication, and marketing sectors. The growth of the internet has seen a profusion of data and a surge in technology for extracting information from big data for marketing strategy, adding value to products and services, and personalizing the consumer experience.