Research On Technical Methods Of Knowledge Service For Product Decision-making Based On Comment Data

The external environment has led to the result that user needs is constantly changing in an increasingly competitive market environment.As a result,enterprises must meet the critical user needs and accelerate the speed of product design updates in order to achieve continuous development in today's highly competitive market environment.Designers usually use traditional methods to obtain user needs information in research,such as expert consultation,user interviews,usability testing,questionnaire methods and so on in the past.However,these methods generally have problems such as lagging needs information sources,high analysis costs,and slowness in capturing needs in the market.In recent years,the Internet has gathered more and more online review data for various products with the rapid development of the Internet.Most of these review data are given by users based on their personal experience with the product.Compared with traditional survey data,the data sources are rich and reliable,easy to obtain,time-sensitive,customer focus enough and so on.Therefore,it can more quickly reflect the needs and preferences of the groups of user for such products and their changes.However,how to effectively dig out more useful requirements and preference information from product review data,and transform it into design knowledge understood by designers is still a problem to be solved.Based on this,the following research that the relationship between the potential value of product review data and product design is carried out:Firstly,this paper obtains user needs based on review mining.Meanwhile,this paper builds a pool of evaluation elements based on the dictionary,Text Rank algorithm and rules to obtain a more accurate calculation of user needs' satisfaction and attention.Further,a screening database for design needs is constructed through the mining of patent and periodical data in order to screen the demand.As a result,the pool of evaluation elements is restructured.Secondly,considering these problems about the traditional needs and its importance information acquisition in the design of the product,this paper takes user needs' satisfaction and attention as the evaluation indicators of needs importance,introduces 2-8 theory,and combines review mining and sentiment analysis technology.On this basis,the research of the importance calculation model of user needs based on comment data is proposed,and the validity of the model is verified by using examples.Then,considering the fact that the information of user needs,needs importance,and so on in the review data cannot be effectively converted into a language that the designer can understand,the QFD model was introduced and the decision-making model of product design based on the QFD that use relatively mature technology to convert the hidden information of the needs and evaluation into design decision-making knowledge understood by the designer based on review data is proposed.Finally,the framework design of the analysis system of decision support based on the theories,methods and technologies mentioned in the paper was carried out,which realized the function of some modules of the system.The system provides designers with decisionmaking knowledge services in the process of the development and design of product.As a result,the cycles of product development is shortened and the efficiency of solving business problems related to product design is accelerated.