Customer Perceived Value Pricing Model Based On Text Analysis And Machine Learning

The emerging economic form,which is based on the Internet platform,is booming,making the form of economic activities more diverse.The gradual maturity of e-commerce platforms has also led more and more consumers to choose the way to buy online.The biggest feature of online consumption is that you can't see the real thing of the goods.You can make consumption decisions by comparing the perception of the value of the goods with the price,and evaluate the purchase behavior afterwards.Such a consumption mechanism provides enterprises with a better opportunity to analyze the feedback of various consumers on commodities,thereby expanding their market share and achieving steady growth in profits.At the same time,how to acquire and use the transaction data of the Internet platform for the development of enterprises is also an urgent problem for enterprise management accounting and operators.Based on the above background,this study explores the quantitative relationship between "customer perceived value"and "electronic consumer goods price",and then derives the product pricing model based on customer perceived value.Different from the traditional customer perception value theory research,this paper rejects the questionnaire survey method with small sample individual error,and uses Internet crawler technology as the source of a large number of customer consumption data.The original unstructured customer review text was used as the data basis for quantifying customer perceived value in this study after the improved participle and feature cutting algorithm in Alistair Kennedy and Diana Inkpen(2010).In order to make the data of this paper closer to the reality of consumer sales,this paper selects the consumption data of electronic consumer goods from Jingdong Mall from 2017 to 2018,mainly considering that the customer perception of the consumer electronics industry is relatively sufficient,and the data structure and data richness can be satisfied to a large extent of the need for design.This study uses the price of products at different times as a marker to distinguish consumer reviews,and more than 200,000 consumer reviews are used as research samples.In order to realize the relationship between multiple independent variables and single dependent variables,this paper will not use the common methods in correlation research,but use the neural network algorithm which is more effective in dealing with multivariate nonlinear data processing to study the research samples.A product pricing model based on customer review-aware data.The results of the study show that customer perception does affect the price of the product to a large extent:this trend is particularly evident in the mid-priced consumer electronics pricing process,which also reflects that the mid-price customer perception is relatively compatible with product pricing.Therefore,the research results of this study provide a relatively practical method for enterprises to analyze the real-time perception of consumers and market sentiment,and then provide reference for enterprises to adjust product prices in time and formulate price adjustment strategies that can maximize the benefits.