System and method for identifying creativity of high-value users in online product community

Abstract

The invention provides a method and a system for identifying creativity of high-value users in an online product community. The method comprises the following steps: S1, extracting eight features representing creativity and user attributes from creativity texts and user-creativity relationships of the online product community; s2, constructing a creative node adjacency matrix; s3, introducing a graph attention neural network, constructing a single attention mechanism, and calculating an attention coefficient alpha ij fusing eight attribute features and an adjacent matrix; s4, constructing a multi-layer attention mechanism, and calculating the accumulated attention coefficient alpha'ij fusing the eight features and the adjacent matrix; s5, constructing a convolutional layer, and outputting a convolutional layer result of a node i; S6, outputting a probability Pi, n of a creative value category corresponding to the node i through a weighted value of a Softmax function (a normalized exponential function), and obtaining a creative value category to which the creative node i belongs; through the steps, a certain creative text proposed by a user is taken as a recognition object, algorithm model parameters are input, and the creative value category and the sorting result of the creative text are obtained.

Description

technical field [0001] Embodiments of the present invention relate to the fields of natural language processing and graph neural networks, and in particular, to a system and method for identifying high-value user ideas in an online product community. Background technique [0002] As an open Internet innovation channel connecting enterprises and users, enterprise-led online product communities have become an important means of gathering product ideas, optimizing user experience, and realizing value co-creation. Product ideas are called high-value user ideas. The online product community not only contains a large amount of original content, but also accumulates a large number of user social relationships, creativity triggers user interaction, and interaction stimulates the emergence of creativity. With the continuous expansion of the product community, how to intelligently identify high-value user ideas based on data + algorithms has become a problem that enterprises are conc...

AI Technical Summary

Problems solved by technology

Existing user idea discovery algorithms have at least two deficiencies: First, the formation and perfection of high-value user ideas is not only the individual behavior of the idea proposer, but also the product of extensive interaction and value co-creation with community users and corporate community administrators. Existing algorithms regard creative content and user comments as isolated and static objects, ignoring the iterative impact of multi-party interaction on user creativity; secondly, in product online communities, the relationship between high-value user ideas and other ideas, and between user creative comments There are dependencies. The existing algorithms ignore the close connection between creative content-user relationship-creative comments. The parameters of the constructed algorithm model depend on the scene, and the robustness and applicability are not high.

Images