Investigating Some Classification Methods To Evaluate Efficiency Results: A Case Study By Using Conjoint Analysis

Abstract

A new product development is an important step of competitive advantage for producers. There are several issues to be considered during developing a new product from the point of view of both customers and producers. Costumer preferences require a great deal of consideration in order to able to address consumer needs in marketing. Conjoint Analysis (CA) is often preferred to reveal utility of the new product by means of customer preferences order on a certain type of product or service which is widely used to reveal how people value different attributes on a new product concept. On the other hand, Data Envelopment Analysis (DEA) can be used to determine efficient product concepts considering both utility and development expenses of the products. In this study, CA was applied with the aim of determining utilities of new car concepts. Then, DEA was used to reveal efficient and inefficient car concepts on a real data set. Finally, most commonly used classification methods Linear Discriminant Analysis (LDA), binary Logistic Regression (LR) and Artificial Neural Networks (ANN) were compared to validate the results of DEA in terms of accuracy.