Information search result generation method and system based on multi-attribute dynamic weight distribution

Abstract

The invention discloses an information search result generation method and system based on multi-attribute dynamic weight distribution. The method mainly comprises the following steps that customer evaluation within a specific time period is inquired, keywords are extracted, evaluation attributes are introduced according to the mentioning frequency, and value type static attributes of commodities are combined to construct multi-attribute sorting options; the different attributes are classified according to the relation between values of the attributes and the satisfaction degree of users, increasing and decreasing functions are used to represent effectiveness functions of two types of attributes respectively, and the effectiveness of each attribute is calculated; a direct evaluation difference-free method or an indirect measuring method is used to select preference of each attribute dynamically, and weight distribution of the attributes is obtained; according to the effectiveness of each attribute and the distribution weight, and a multi-attribute effectiveness function is used to calculate the effectiveness score of each commodity, and the commodities are ordered according to the effectiveness scores.

Description

Technical field [0001] The invention relates to a method and system for generating information search results based on multi-attribute dynamic weight distribution. Background technique [0002] In e-commerce, the sorting function plays an important role in helping users to purchase suitable products. Through the sorting function, the products are sorted in ascending / descending order of a certain attribute; the products with the best value of the attribute are ranked first, such as the lowest price or the best-selling. The attributes in the ranking can be roughly divided into two categories: numeric static attributes (that is, the price and specifications of the product) and customer reviews. [0003] Generally, customer evaluation refers to the overall rating of the product by the customer. However, the overall score is often not effective in helping customers find suitable products. For example, a certain attribute is important to customer A, but may not matter to customer B. T...

AI Technical Summary

Problems solved by technology

[0006] 1) The calculation process is complicated, and a large number of repeated calculations are required for each person, different keywords, and historical preferences, causing the system to run slowly and prolong the waiting time;

[0007] 2) It cannot be associated with inherent or commonly used e-commerce platforms, and the scope of application is very narrow and cannot be promoted;

[0008] 3) For a large number of keywords or preferences input at one time, calculation errors are prone to occur in the process of establishing a collection and screening, resulting in no recommended results or situations where the recommended results do not match the input

[0009] Therefore, the traditional information search and sorting methods in the existing e-commerce can no longer give users a good experience.

Images