User Needs Mining For Online Reviews Based On KANO Model

Customized production makes a large number of companies pay more attention to the exploration of users’ potential needs.A large number of online reviews generated along with e-commerce and social networking sites,becoming an emerging data source for companies to tap user preferences and needs.Online reviews which have rich semantics are actively expressed by users describes user needs more objectively.so how to effectively mine the needs contained in reviews has attracted widespread attention from the academic community and the industry.This task not only requires the use of in-depth mining methods to process online reviews,but also requires the integration of valuable data brought by online reviews with needs analysis models.However,current research focuses more on the previous part,mainly using feature extraction,sentiment analysis and other methods to mine needs in online reviews.These researches have neglected the integration with the user demand model,and lacked the construction of a complete process from online comment mining to user demand analysis.It is difficult to further process and analyze the basic demand information mined from online reviews.By systematically combing the research on online reviews and user needs,a KANO needs analysis framework based on aspect-level sentiment analysis for online reviews is proposed.The needs analysis framework is roughly divided into two parts.The first part is to use the aspect level sentiment analysis method to excavate the user’s multi-faceted emotional tendency.It mainly includes four steps:online comment data preprocessing,product attribute extraction,construction of comment segmentation model based on BERT pre-training model,and construction of sentiment classification model based on BERT pre-training model.Secondly,analyzing user needs and satisfaction through the KANO model.The second part analyzes the combination of user demand and satisfaction degree through Kano model.It calculates user satisfaction(STF_i)and user attention(PFC_i)based on the basic concepts of the KANO model.Then,the user needs analysis of the KANO model based on online reviews is realized.Based on this analysis framework,the relevant online comment data under the category of "mobile phone"obtained from"Jingdong Mall" is used to complete attribute extraction,review segmentation,sentiment analysis and KANO demand analysis to verify the effectiveness of the model.Experimental results confirm that the analysis framework can effectively extract information from online reviews.It can realize user demand analysis under the attributes of mobile phone products,and provide decision-making suggestions for related manufacturers.The demand analysis framework for online reviews proposed in this paper can completely realize the whole process from data mining to demand analysis.The demand analysis framework for online reviews proposed in this paper can completely realize the whole process from data mining to demand analysis.At the practical level,the demand analysis framework provided in this paper can help companies further obtain the potential needs of users and pay attention to the generation and changes of user needs in a timely manner.For different types of needs,companies can make decision-making adjustments based on their actual conditions,so as to achieve the effect of enabling companies to improve product quality,increase users,and enhance word-of-mouth.