Research On The Ranking Of Electronic Technology Products Based On Multi-source Text Data Under Fuzzy Environment

With the continuous improvement of the income level of Chinese residents and the continuous development of science and technology,the demand of Chinese residents for consumer electronic technology products represented by sweeping robots and mobile phones is also increasing.Electronic science and technology products are characterized by high price,strong technical attributes,complex software and hardware parameters,so most consumers will actively seek various reference ways to obtain relevant technical information and comparative evaluation information of target electronic science and technology products before purchasing electronic science and technology products.The online review text data of e-commerce platform and the evaluation long text data of online content community platform are the main reference sources.However,for most consumers,it is difficult to have a comprehensive understanding of electronic technology products due to the lack of corresponding technical knowledge reserves and high learning costs,and then it is difficult to make the best choice quickly among the target electronic technology products of the same kind,which will finally hit consumers’ purchase intention.However,if we can provide a real-time comparison and ranking of similar target electronic technology products according to the needs of consumers,and provide consumers with a scientific reference,it will greatly reduce the learning costs of consumers,and then promote the purchase behavior of consumers,which has a high realistic value.However,most of the existing research on product ranking only relies on a single online evaluation information.However,when the product being ranked is an electronic technology product,due to the limitations of most consumers’ cognition of such products and the bias in their perception of the importance of product attributes,online reviews of electronic technology products may have issues such as attribute confusion and easy omission of attributes,Therefore,relying solely on a single online review text mining to obtain product attribute results can lead to major issues such as incomplete attributes and difficult classification of attribute feature words,which can lead to inaccurate aspect level emotional orientation analysis results,and the resulting product ranking results may not be consistent with the actual situation.In addition,when considering attribute weights,most studies use fixed values to represent product attribute weights.However,in real life,decision makers may experience hesitation and group inconsistency when judging the importance of attributes.Therefore,using fixed values to represent attribute weights cannot reflect the reality.Based on the above practical background and research deficiencies,this thesis puts forward a new idea of sorting electronic technology products based on multi-source text data in fuzzy environment.The data source is not only the short text of product online review on ecommerce platform,but also the long text of product evaluation on content community online platform.The product attributes are determined by the intersection of these two text data,the product evaluation matrix is constructed by the short text data of online review,and the product attribute weight is determined by the long text data of product evaluation.The evaluation value and attribute weight value are expressed by intuitionistic fuzzy number and probability hesitation fuzzy number respectively.At the same time,the intuitionistic fuzzy concurrent integration operator(PD-IFUIA)with the weight of probabilistic hesitation fuzzy number and the improved probabilistic dual hesitation fuzzy entropy are innovatively proposed to solve the intuitionistic fuzzy multi-attribute decision-making problem with the attribute weight of probabilistic hesitation fuzzy number.The main research contents and innovative achievements of this thesis are as follows:(1)Aiming at the disadvantages of sorting electronic technology products based on only a single online review short text data,this thesis proposes a new idea of sorting electronic technology products based on multi-source text data in a fuzzy environment.The combination of long and short text data for product attribute mining ensures the comprehensiveness and professionalism of the product attribute set.The improved BERT model is used to analyze the emotional tendency of the online comment short text data that directly reflects the consumer’s own experience and feelings,and the product intuitionistic fuzzy evaluation matrix is constructed according to the analysis results,and the product attribute probability hesitation fuzzy weight is determined based on the more professional evaluation long text data,Make the attribute weight information more reliable and more realistic.(2)A product ranking method based on the new PD-IFUIA integration operator is constructed.Based on the theory of probabilistic dual hesitant fuzzy sets and intuitionistic fuzzy sets,this thesis constructs the intuitionistic fuzzy concurrent integration operator(PD-IFUIA)with the weight of probabilistic hesitant fuzzy numbers,and proves the monotonicity and boundedness of the integration operator,thus ensuring the applicability of the integration operator to intuitionistic fuzzy multi-attribute decision-making with the weight of probabilistic hesitant fuzzy numbers,The score value of each alternative product is calculated by combining PD-IFUIA operator and improved score function,and the alternative products are sorted according to the size of the score function.(3)A product TOPSIS ranking method based on improved probability dual hesitation fuzzy entropy is constructed.Based on the relevant theories of probabilistic hesitation fuzzy entropy and probabilistic dual hesitation fuzzy entropy,this thesis improves the probabilistic dual hesitation fuzzy entropy and proves its relevant properties,which makes up for the deficiency of the existing probabilistic dual hesitation fuzzy entropy that only considers the degree of hesitation but not the degree of ambiguity.Based on the new probabilistic dual hesitation fuzzy entropy,the comprehensive correlation coefficients between each option and the positive and negative ideal options are calculated,and the proximity degree of each alternative product option is calculated based on the comprehensive correlation coefficient.The alternative products are ranked based on the proximity degree.