Open AccessINTEGRATED APPROACH OF FUZZY MULTI-ATTRIBUTE DECISION MAKING AND DATA MINING FOR CUSTOMER SEGMENTATION
Author(s) -
Manidatta Ray,
Mamata Ray,
Kamalakanta Muduli,
Audrius Banaitis,
Anil Kumar
Publication year - 2021
Publication title -
e+m. ekonomie a management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.318
H-Index - 22
eISSN - 2336-5064
pISSN - 1212-3609
DOI - 10.15240/tul/001/2021-4-011
Subject(s) - topsis , ideal solution , data mining , fuzzy logic , computer science , analytic hierarchy process , ranking (information retrieval) , cluster analysis , customer relationship management , operations research , artificial intelligence , mathematics , database , physics , thermodynamics
This research work focuses on integrating the multi attribute decision making with data mining in a fuzzy decision environment for customer relationship management. The main objective is to analyse the relation between multi attribute decision making and data mining considering a complex problem of ordering customers segments, which is based on four criteria of customer’s life time value, viz. length (L), recency (R), frequency (F) and monetary value (M). The proposed integrated approach involves fuzzy C-means (FCM) cluster analysis as data mining tool. The experiment conducted using MATLAB 12.0 for identifying eight clusters of customers. The two multi attribute decision making tools i.e., fuzzy AHP (Analytic Hierarchy Process) and fuzzy TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) are used for ranking these identified clusters. The applicability of the integrated decision making technique is also demonstrated in this paper considering the case of Indian retail sector. This research collected responses from nine experts from Indian retail industry regarding their perception of relative importance of four criteria of customer life value and evaluated weights of each criterion using fuzzy AHP. Transaction data of 18 months of the case retail store was analysed to segment 1,600 customers into eight clusters using fuzzy c-means clustering analysis technique. Finally, these eight clusters were ranked using fuzzy TOPSIS (Technique for Order Preference by Similarity to Ideal Solution). The findings of this research could be helpful for firms in identifying the more valuable customers for them and allocate more resources to satisfy them. The findings will be also helpful in developing different loyalty program strategies for customers of different clusters.