In recent years, with the rapid development of the e-commerce industry, there has been an artificial manipulation of product reviews on many e-commerce platforms, making it difficult for consumers to distinguish the quality of products. Therefore, identifying fake reviews can promote the high-quality development of the e-commerce industry under new development patterns. This article collects comment data from a certain e-commerce platform, selects text similarity, richness, length, number of pictures and videos, and emotional intensity as relevant indicators of fake reviews based on existing research results at home and abroad, combined with the comment data collected in this article. Based on the collected data set, the article extracts three key variables from the comment text: the text data that only contains the initial comments, the text data that contains the appended comment, and the text data that contains both initial and additional comments. Then, the article uses the entropy weight method to calculate the weight of each indicator, and calculates the comprehensive score of each comment by using the TOPSIS model. Finally, by combining the comprehensive score with the corresponding comment text for analysis, the conclusion drawn in this article is: the proportion of comment data that only contains initial comments is relatively high in product reviews, and there are many fake reviews among them; the proportion of appended comments is relatively low in e-commerce website reviews, but the authenticity is high, and the degree of fakeness is relatively low.