Quality Improvement Strategies of Mobile Phone Product Based on Text and Sentiment Analysis

e-commerce reviews to build a new sentiment lexicon, combined with python language to traverse degree adverbs, negation words, etc., to calculate the sentiment value of each review sentence, so as to realize the sentiment tendency classification of e-commerce products. Zul Abstract: The development of the Internet has brought people the opportunity to communicate online, and user reviews have appeared under the review interface of various brands of hot products, and these reviews containing emotions have also intensified the competition among products. This paper takes cell phone background as an example, through data mining cell phone user reviews, using word frequency and word cloud methods for text analysis of user reviews, extracting two types of hot review words-[service features] and [cell phone features], exploring whether each hot review word behind can significantly affect the favorable rating of cell phones by establishing a random forest regression model, and calculating the corresponding sentiment score of each hot review word according to the sentiment tendency calculation method. Finally, this paper analyzes each hot review term in depth and designs a specific evaluation system, which can be used to paint an overall and detailed portrait of the cell phone, thus helping merchants to quickly find the shortcomings of the product and determine the direction of improvement.