Research On Style Recommendation System Of Sports Leisure Coat Based On Consumer Style Preference

In the era of "emotional consumption and experience economy," consumers’ physiological needs have been met,and they place more emphasis on psychological feelings and tend to choose clothing styles that match their own preferences.However,clothing style definitions are often vague and difficult to control,creating a barrier between designers and consumers.The abundance of data makes it difficult for consumers to quickly select clothing that meets their needs,which contradicts today’s fast-paced lifestyle.Therefore,quantifying the difficult-to-qualify clothing styles can help meet consumer needs,establish a communication bridge between designers and consumers,and combined with personalized recommendation systems,can improve consumer purchasing efficiency and satisfaction.Influenced by health concepts and social environments,the sportswear market is on the rise.This article takes sport jackets as an example to build a quantification model to predict consumer preferences and provide personalized recommendations.The main research contents are as follows:Firstly,the classification standard of style and perceptual quantitative method are studied,so as to lay a theoretical foundation.The modular idea is further used to divide the style attribute elements,and the hierarchical structure chart of style design elements is constructed by combining the images downloaded by web crawler technology and the key attributes screened by expert interviews.On this basis,the AHP is used to construct the evaluation model of style elements and assign corresponding weight to the attributes.Secondly,in order to intuitively reflect the transformation law of style with style,choose to present the relevant content in the form of charts.The clustering method was used to select 32 representative samples.After the samples were standardized,the perception difference was reduced by combining the two-dimensional human platform.On the basis of reaching the reliability standard,the broken line chart of perceptual image score of samples was drawn to explore the image distribution law.Combined with correlation analysis,principal component analysis and other methods,three factors were extracted as representative factors of style perception,three-dimensional image space and two-dimensional planar distribution map were constructed,and the correlation between style attributes and style perception was analyzed.Then,the style items and categories were quantified and coded using the quantitative theory I as input variables,and the influence of style on perception factors was analyzed by combining the output variable vector to build a style quantitative model.The prediction accuracy and fitting degree of the model were tested by negative correlation coefficient,decision coefficient and Durbin-Watson,and the effect of the model was verified in combination with actual cases to provide a basis for style recommendation.Finally,the evaluation model and quantitative model are applied to the recommendation system.Based on the multiple demands of consumers,the recommendation strategy was formulated,the multi-attribute decision method was used to provide users with a style constraint scheme,the Top-N principle was used to visually display 7 clothing types and their serial number,common and unique attributes,and the dynamic preferences of consumers were predicted by combining user feedback and the theory of multiple alternative decision field.Ten consumers are randomly selected for satisfaction test,and the overall utility value of the recommendation system is calculated.It can be seen that the recommendation effect has reached a satisfactory level.