Product Success Evaluation Model Based on Star Ratings, Reviews and Product Popularity

In the era of the Internet and big data, it is of great significance to formulate product sales strategies and product improvement plans based on user evaluation data. Therefore, an evaluation system is established to calculate the index (product success index) composed of three factors: actual star rating of consumers, star rating of review analysis, and star rating of product popularity (star rating of products over time). We use AHP to determine the weights between these three factors. We then select the products we consider most successful based on the calculated product success index and found that users are most concerned about product quality, price and security. At the same time, to solve the question of whether a particular review affects subsequent ratings. We used a two-way ANOVA method to analyze whether specific reviews in the early stages would affect the stars in later stages. We found that the p-value is concentrated between 0.1-0.2, which means the specific star rating is considered to have an impact on the evaluation. Finally, we suggest that Merchants and platforms should pay more attention to the characteristics of successful products and early product feedback. The innovation of this paper is that it combines a variety of methods including systems science, management science, and statistical prediction artificial intelligence to comprehensively evaluate and predict the success of commodities, which can provide certain support for business decisions.