Evaluating the Value of User Comments on the Platform based on K-means Clustering

For today’s businesses in e-commerce platforms, profiting by classifying the value of user comments, identifying important value comments and improving goods according to them are helpful to understand user’s purchase preferences and accurately improve products that meet user expectations, which can not only improve user satisfaction but also reduce marketing cost. This paper presents a comment value classification method based on K-means clustering. First, this method improves the RFM model of customer value, constructs the RFM-RES model suitable for comment value classification, and standardizes the index system. After that, analyzes examples of user comments on "DaZhongDianPing" (a review platform) using the presented model, uses the K-means method based on Elbow to cluster the processed comment data set, and completes the initial screening of the comment value, then confirmes the accuracy of the method by comparing the consistency of the score of the comment group of each category with the actual score. Finally, the reference value of each category is analyzed according to the evaluation results. Using this method for evaluation platform comment value classification, under the premise of ensuring reference value, can greatly improve the efficiency of decision-making management to screen effective comments for product improvement.