Asia Pacific University Library catalogue

A thesis submitted in fulfillment of the requirement for the award of the degree of M.Sc in Data Science and Business Analytics (UCMP1606DSBA)

Market segmentation is one of the core components of marketing strategies to determine the success rate of a business. The primary goal of creating an effective market segmentation is to profile the audience according to their interpersonal background, needs and preferences. Therefore, the potential group of people will be targeted to customize products offerings, personalize marketing communication, increase revenue which leads to resources saving such as money and time investment. This study will emphasize on direct segmentation and behavioral targeting. Direct segmentation leverage on consumer attributes including demographic, geographic and psycho-graphic to fulfill their exact needs. Whereby, in direct segmentation, consumers are also monitored in aspect of behavior response factors and socio-economic status for product as well as communication targeting. Consumers' interest factors will be triggered to respond to an offering instead of fulfilling their needs. Nevertheless, there are several underlying problems with market segmentation when they are composited without complete knowledge and profiles are built based on general descriptions. This study aims to propose prediction models to identify customer spending and movie preferences based on interpersonal behavior. The factors such as gender differences and identification of unique trends among consumers were considered in the segmentation through exploratory data analysis. This approach is highly essential to plan niche marketing which specifically focus on a group of audience based on their unique qualities that has been explored. A real-world survey dataset with 150 variables and 1010 records have been used for this study.. Spending behavior of audiences were intended to be predict based on music interest. Hence, the music genres were clustered using Cubic Clustering Criterion (CCC) to identify the most prominent music category. Seven clusters were formed in total. The clusters were predicted holistically and individually merely to produce more customized products to trigger product purchase without much hesitation. In addition, consumers were also noted to be actively inclined towards,movies, as such, hobbies and interest factors were predicted to assess the level of contribution towards movies preferences. Linear Regression, Auto Neural Network and Neural Network were developed and compared to opt the best model. Auto Neural Network provided the best prediction result with least average squared error of 0.003257. The study concludes that market segmentation is essential for niche marketing. The fundamental focus for segmentation should not only be generalized towards the common attributes but also unique interpersonal values.