An integrated HPF-TODIM-MULTIMOORA approach for car selection through online reviews

The fierce competition in automobile industry makes buying a car to be a daunting task for customers, since it is a high-risk decision due to the intrinsic imperceptibility of car products. Nowadays, customers are more willing to take some advice from online reviews before making a purchase decision in the era of big data. Therefore, we try to develop a novel data-driven method to make reasonable decisions for customers when they buy cars. To achieve the goal, our study first collects data from online reviews rather than questionnaires. To better depict the ambiguity and complexity of online reviews, we then utilize the hesitant probabilistic fuzzy set (HPFS) and sentiment analysis to quantify evaluations. Taking customers’ psychological cognition into account, a novel method combining the prospect theory and term frequency is used to measure the weights of different attributes. After that, we further propose an improved multiplicative multi-objective optimization by ratio analysis (MULTIMOORA) approach to rank the products. Specifically, the dominance degrees from TODIM (an acronym in Portuguese of interactive and multicriteria decision making) method are extended in three sub-ranking methods in MULTIMOORA, namely ratio system method, reference point method, and full multiplicative form method to prioritize all car products for customers. Finally, we provide some product improvement suggestions for car manufacturers.