UNDERSTANDING CONSUMER PREFERENCE THROUGH FUZZY-BASED RECOMMENDATION SYSTEM

We introduce a new recommendation technique that incorporates the general consumer perception about product attributes while elucidating buyers’ trade-offs amongst the attributes of a multi-attributed product such as the smartphone, during purchase decision-making. The concepts of 2-tuple fuzzy linguistic representation and fuzzy market research system (FUZMAR) approach have been used to develop this recommendation technique. In this paper, the power of an attribute in predicting another, especially in fuzzy and uncertain environments, has been explained through the conceptualisation of consumers’ innate associations between the attributes. The predictive powers of the attributes are used to derive the weights associated with the attributes, using analytical hierarchy process, which in turn gives consumers’ sequential preference of products. The proposed recommendation technique is illustrated with an empirical study. 

A qualitative study was conducted to understand how consumers make smartphone buying decisions, the attributes that are most important for consumers, and to capture respondents’ verbatim responses about attributes of suggested smartphones. The study revealed that the respondents considered a few product attributes to be more important than other attributes while making purchase decisions. Also, it was discovered that respondents visualise interactions amongst these attributes. They weigh one attribute based on the knowledge of other attributes. Taking respondents’ perception on these attributes (numerical and linguistic data) as input, we captured the respondents’ overall ordered preference amongst the selected smartphones through our recommendation technique. Online recommendation sites can leverage from this technique by designing their web portals according to the consumers’ choice behaviour and facilitating better online search experiences for consumers.