Research On Influencing Factors Of Agricultural Product Consumer Satisfaction

In the information times of internet +,with the continuous development of the e-commerce platform for agricultural products and rapid trade volume,consumer satisfaction with products or services is critical to the development of businesses and e-commerce platforms.There is relatively less current research on consumer satisfaction based on reviews,especially for agricultural products.In addition,most research methods and model acquisition methods are difficult to measure consumer's true emotional attitude objectively and effectively.Therefore,a reasonable quantification of consumer sentiment about product features is an effective way to deeply explore the factors that influence consumer satisfaction.The paper uses JAVA language to design crawler program to collect rice online commentary data;combines semantic analysis technology and sentiment analysis methods to quantify product feature emotion value;introduces text sentiment analysis technology to construct improved CCSI consumer satisfaction model;and applies multiple regression to analyze the relationship between product characteristics and overall consumer satisfaction.The above research provides a theoretical basis for the promotion of marketing strategies for product vendors and e-commerce platforms.The paper launches innovative research mainly from the following aspects:(1)In order to make effective use of the review text data and deeply tap the consumer satisfaction,the paper reconstructs three major emotional dictionaries of HowNet,NTUSD,and Chinese vocabulary.Using the improved PMI algorithm and combining the manually annotated emotions,the paper builds a basic sentiment dictionary to provide a basis for the reasonable quantification of feature evaluation units in the review text.(2)Research the method of extracting vocabulary used to describe product feature and calculating emotion value.Firstly,the appraisal expression is identified based on the function module of word segmentation,part of speech tagging and semantic analysis in the LTP language technology platform.At the same time,the JAVA language is used to parse the XML result document and extract feature evaluation unit and sentiment evaluation unit.Secondly,combined with the syntactic dependency of the evaluation unit and the tagged part of speech,the emotion value calculation rules are designed to quantify the emotional attitudes carried by the evaluation unit,and the quantified results are used as explanatory variable data in the CSI model.(3)Based on the traditional CCSI model,a consumer satisfaction index model is sto study therelationship between pre-variables and consumer satisfaction.At the same time,in order to compress the result variables and interpretation variable dimension,the paper uses min-max the method to standardize the model data.(4)The paper uses statistical methods to study the impact of product features excavated in the review text on overall consumer satisfaction,and researches the impact weight of each feature in overall satisfaction.The most significant factor affecting rice consumer satisfaction is Taste and service,followed by brand,appearance and logistics,and the price is the minimum factor.