Sequence recommendation method and system based on dynamic interaction attention mechanism

A technology of dynamic interaction and recommendation method, which is used in sales/rental transactions, instruments, electronic digital data processing, etc., to achieve the effect of accurate recommendation results

Active Publication Date: 2019-09-17
NAT UNIV OF DEFENSE TECH
View PDF14 Cites 29 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to provide a sequence recommendation method and system based on a dynamic interactive attention mechanism, so as to solve the technical problems existing in the prior art

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Sequence recommendation method and system based on dynamic interaction attention mechanism
  • Sequence recommendation method and system based on dynamic interaction attention mechanism
  • Sequence recommendation method and system based on dynamic interaction attention mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0053]Sequence recommendation is to predict the next object that the user may be interested in based on the user's interaction history over a period of time. A reasonable and effective sequential recommendation model should be able to effectively combine users' long-term and short-term preferences. Existing methods always regard the user's long-term interest preference as a static and fixed vector, and the user's long-term preference remains unchanged when dealing with different dynamic preferences of the user. But this is unreasonable, because in the face of users' different dynamic preferences, their long-term preferences should have different importance. The relative importance of events in a user's long-term interaction history depends on the events in his short-term interaction history, and vice versa. If a user has searched for cameras in the current session, long-term electronics-related interactions in the user's long-term search history should be more important than ...

Embodiment 2

[0108] In this embodiment, the Tmall e-commerce data set and the Tianchi e-commerce data set are selected, and the specific information of the data sets is shown in Table 1 below:

[0109] Table 1

[0110]

[0111]

[0112] This experiment selected the following models as comparison objects: Item-pop is a method of sorting items according to the number of interactions of items, and then recommending according to the ranking, which is a non-personalized method. FPMC is a recommendation method based on Markov chain and collaborative filtering. GRU4Rec, a recommendation model based on RNN and query sessions, is a non-personalized approach. HRNN is used for personalized query-session-based recommendation method, which adopts a hierarchical RNN-structure, namely session-level RNN and user-level RNN, which are used to simulate short-term and long-term preferences of users. NARM is based on the RNN model and applies the attention mechanism to capture the user's preference inf...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a sequence recommendation method based on a dynamic interaction attention mechanism. The method comprises the steps of obtaining initial short-term preference and initial long-term preference of a user; obtaining long-term preferences and short-term preferences according to the interactive attention network in combination with the initial short-term preferences and the initial long-term preferences; and scoring the corresponding articles according to the sequence recommendation model in combination with the long-term preference and the short-term preference, and recommending the articles to the user according to the scoring result. According to the constructed dynamic interaction attention mechanism network model for sequential recommendation (DCN-SR), long-term and short-term interaction joint dependence expression of the user can be learned through the model, and the recommendation result is more accurate in combination with long-term preference and short-term preference.

Description

technical field [0001] The invention belongs to the field of sequence recommendation, and in particular relates to a sequence recommendation method and system based on a dynamic interactive attention mechanism. Background technique [0002] Recommender systems are an effective solution to help people cope with the increasingly complex information environment. Traditional recommender systems usually ignore sequential information and focus on mining static correlations between users and items from interactions. For example, a typical traditional recommender system based on matrix factorization can learn from the entire interaction history to effectively model users' general preferences, but it does not model users' short-term sequential interaction behavior. Different from traditional recommendation systems, sequence recommendation is to predict the next object that the user may be interested in based on the user's interaction history over a period of time. [0003] Existing...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/9535G06F16/2458G06Q30/06
CPCG06F16/9535G06F16/2474G06Q30/0631
Inventor 蔡飞陈洪辉刘俊先罗爱民舒振陈涛罗雪山
Owner NAT UNIV OF DEFENSE TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap