Abstract: In order to support consumers’ purchasing decision， a product purchasing recommendation ranking method based on multi-attribute online ratings information was proposed. First， the online rating information on each product attribute was transformed into the probability distributions of attribute rating scales， and the cumulative distribution functions of online evaluation results to each product attribute were obtained. Then， the weighted cumulative distribution functions decision matrix was constructed， and the weighted cumulative distribution vectors of the ideal and anti-ideal products were determined based on the decision matrix. Furthermore， the distances between the weighted cumulative distribution vectors of each alternative product and the ideal/anti-ideal product were calculated， and the closeness coefficient of each alternative product was obtained. According to the closeness coefficient， the recommendation ranking results of alternative products were determined. Finally， an example to support consumers’ car purchasing decision was used to illustrate the feasibility and validity of the proposed method.

Key words: product purchasing decision, online ratings information, probabilistic distribution, TOPSIS(technique for order preference by similarity to an ideal solution), recommendation ranking