Analysis on Preference of Citrus Attributes Using K-MEANS Clustering Analysis and Latent Class Analysis

This study aims to understand consumer preferences for citrus by segmenting them based on their attributes according to Lancaster's consumer theory. Data was collected from 200 respondents through a choice-based conjoint survey that included four attributes of citrus: weight, size, price, and packaging. Consumer preferences were clustered using two techniques: K-MEANS clustering and latent class analysis. Individual utility coefficients and willingness-to-pay (WTP) were estimated using a mixed logit model before conducting K-MEANS cluster analysis, and principal component analysis (PCA) was used to reduce dimensionality. The results of the two cluster showed some similarities but shared limitd information. K-means cluster analysis is advantageous in identifying distinct consumer groups by simultaneously considering preferences for multiple attributes, while LCA can be used to analyze consumer preferences for a wider range of attribute combinations in more detail, which is advantageous for developing customized marketing strategies targeting specific consumer groups. The results of this study can be used to better understand consumer preferences in the citrus fruit market and to develop effective marketing strategies.