Development of Hybrid Recommendation Model for Online Shopping Based on Machine Learning

With the development of cloud computing, big data and other information technologies, e-commerce platforms have rapidly emerged as the main front of consumer transactions, online reviews have shown explosive growth. Online review data contains lots of valuable information, through text mining techniques to analyze this information, to help the company design more attractive products for online sales and develop a more complete marketing strategy. In this paper, the data of online shopping reviews, star ratings and “helpfulness ratings” published by consumers on Amazon are used as the research object. Firstly, based on the sentiment analysis of commodity evaluation text, we can obtain the average sentiment intensity of each product’s features. Secondly, the utility index is defined by considering time factor and reviews’ values, with the calculated results, K-means clustering method is adopted to classify reviews and identify significant features of informative reviews. Finally, through lasso algorithm, this paper filter out influential words and their influence coefficients. Sensitivity analysis have shown that the model is significant and robust.