A Comparative Analysis of the CRITIC and Entropy Methods for Objective Weighting of Priority Criteria

Abstract

Various criteria weighting methods are available, and this study aims to compare the Criteria Importance Through Inter Criteria Correlation (CRITIC) and Entropy methods to determine the criteria weights. This case study focuses on identifying priority customers from 2 years of sales transactions in an online retail company, which processes more than 1 million transactions with 8 features. The researcher selected 100 high-value customers as alternative data, prioritizing research efficiency and high-value insights. Four criteria were set for customer prioritization. Sensitivity analysis was conducted using the Additive Ratio Assessment (ARAS) method to measure the stability of the method. The CRITIC method produced balanced weights (0.23-0.27), while Entropy produced more variable weights, with C3 being the largest criterion weight with a value of 0.46, indicating its strong dependence on the data distribution. Sensitivity analysis revealed that the Entropy-ARAS method was more sensitive to weight changes (75.11134%) in this customer prioritization case compared to the CRITIC-ARAS method (56.95372%).