Research On Product Recommendation Ranking Of E-commerce Platform Based On Quality Function Development

In the context of e-commerce and big data era,consumers are more concerned about the consumption experience of online shopping.Due to the wide variety of goods on the online platform and the varying quality of goods,consumers need to spend more time and energy in comparing and selecting products.At the same time,the untouchability of online products makes online shopping users can only judge whether the products meet their expectations through the information on the Internet,which increases consumers’ purchase risk and reduces their desire to buy.In addition,online review data is usually unstructured text data,and its large volume and small amount of key information increase the time cost for consumers to select products,which may reduce consumers’ shopping enthusiasm.Thus,the study of the product recommendation ranking problem on e-commerce platforms plays an important role in reducing consumer decision costs and improving consumer satisfaction.On the other hand,for the third-party e-commerce platform that connects production suppliers and consumers in the supply chain,product recommendation ranking provides a reference for the selection of product suppliers by the dimensional platform while improving sales efficiency.This paper summarizes the current research status of online review and information mining,product recommendation ranking,quality house and user portrait issues by scholars at home and abroad around data mining,quality function development,user portrait and other related theories,and points out that the current research on using quality house to measure the matching degree between products and consumers to achieve e-commerce product recommendation ranking is relatively scarce,so as to propose the research of product recommendation ranking based on house of quality for e-commerce platform.The author first uses data mining technology to process the review information obtained,and extracts product feature words by using TextRank algorithm,then constructs the spatial vector of consumer product attributes by using sentiment tendency analysis,and divides customer groups by combining with improved two-step clustering algorithm,so as to fill the "consumer-product attribute" and "product-product attribute" house of quality models.In addition,this paper proposes to further optimize the original model based on multidimensional labeling,taking into account the basic attributes,purchasing habits and behavioral characteristics of consumers.After six user portrait labels are selected from two aspects of basic consumer attributes and consumer consumption characteristics,sample data are collected and quantified through sampling and survey interviews,and the obtained multidimensional vectors of consumer labels are divided into consumer groups using an improved two-step clustering algorithm,which are also brought into the quality house model to optimize the original product recommendation ranking method.In addition,this article verifies the feasibility and usefulness of the product recommendation ranking method based on house of quality and the product recommendation optimization method based on multidimensional labeling with the help of lipstick consumer review and survey interview information from Tmall Platform.