Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Influence maximization method in social network and based on theme interest

A technology of maximizing influence and social network, applied in special data processing applications, instruments, electrical digital data processing, etc., it can solve the problem of not considering the characteristics of propagation items, the TIC model does not take into account the distribution of user interests, and cannot accurately describe information propagation. issues of laws

Active Publication Date: 2018-06-22
江苏派智信息科技有限公司
View PDF4 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0031] The purpose of the present invention is to solve the problem that the existing IC model does not consider the characteristics of the dissemination items, assuming that the impact probability on all dissemination items is the same; and the existing TIC model does not take into account the user's interest distribution, and cannot accurately describe information dissemination Laws, leading to the problem of low prediction accuracy of information dissemination, and a method of maximizing influence based on topic interests in social networks is proposed

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
  • Influence maximization method in social network and based on theme interest
  • Influence maximization method in social network and based on theme interest
  • Influence maximization method in social network and based on theme interest

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0081] Specific implementation mode one: combine figure 1 Describe this embodiment, the specific process of the influence maximization method based on topic interest in the social network of this embodiment is:

[0082] Step 1. Establish a communication model TI-IC based on topic interest;

[0083] The present invention introduces the topic interest based propagation model TI-IC. The TI-IC model is an extension of the IC model, assuming that there is a topic distribution for each communication item and an interest distribution for each user. The present invention introduces the topic interest based propagation model TI-IC. The TI-IC model is an extension of the IC model to mix topics in each spread item and user. For example, a movie may contain the following basic themes: comedy, love, action, etc., and a user also has an interest distribution, such as 0.6 for comedy, 0.1 for romance, and 0.1 for action movies. The degree of liking is 0.3.

[0084] Given a social network...

specific Embodiment approach 2

[0148] Specific embodiment two: what this embodiment is different from specific embodiment one is: utilize EM algorithm to learn the parameter of TI-IC model in described step 2, obtain the output of EM learning algorithm, namely the parameter Θ of TI-IC model, Θ include with The specific process is:

[0149] Step a) Initialize with a normal distribution with a mean of 0 and a variance of 0.01 π z ; z ∈ [1, Z], u ∈ V;

[0150] π z Indicates the prior probability of all propagation items i on topic z;

[0151] Step b) For all propagation items i and topics z, calculate (EM algorithm E step);

[0152] Step c) For all topics z, calculate (EM algorithm M steps)

[0153] Step d) For all topics z and users u, calculate (EM algorithm M steps);

[0154] Step e) repeatedly execute step b) to step d), until convergence;

[0155] Step f) output with

[0156] That is, the output Θ of the EM learning algorithm is obtained, and Θ includes with

[0157] The pseudo...

specific Embodiment approach 3

[0164] Specific implementation mode three: the difference between this implementation mode and specific implementation mode one or two is: when a new propagation item i appears in the step 2, the topic distribution vector of the new propagation item i is solved The specific process is:

[0165] When a new propagation item i appears, the following method is used to solve the topic distribution vector of the new propagation item i Minimize the following objectives;

[0166]

[0167] For the above optimization objective, use the gradient descent algorithm to solve; The specific solution process is:

[0168] Step S1) Initialize with a normal distribution with a mean of 0 and a variance of 0.01

[0169] Step S2)

[0170] Step S3) repeatedly execute step S2), until convergence;

[0171] Step S4) output

[0172] where λ is the learning step size;

[0173] The specific solution process is shown in Algorithm 2, Algorithm 2: Algorithm for learning new propagation it...

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 relates to influence maximization methods of theme interest, in particular to an influence maximization method in a social network and based on theme interest, and aims to solve the problem of low information dissemination prediction accuracy caused by the fact that existing IC models do not take features of dissemination items into consideration and suppose probability of influenceon all dissemination items to be equal and existing TIC models do not take interest distribution of users into consideration and cannot accurately describe information dissemination laws. The method includes: step 1, building a dissemination model TI-IC based on theme interest; step 2, utilizing an EM algorithm to learn parameters of the TI-IC model and theme distribution vector of a new dissemination item; step 3, putting forward an influence maximization algorithm for the TI-IC model based on the step 2. The method is used in the field of influence maximization problems of the social network.

Description

technical field [0001] The present invention relates to methods of maximizing the impact of subject interests. Background technique [0002] In recent years, with the popularity of social applications, the way people obtain information has undergone great changes. Forwarding and sharing news through online social networks has gradually become the main way for people to obtain information. Many online social networking sites allow users to retweet, comment, flag, or perform similar operations on information. If we can fully mine these massive data in social networks and discover the rules of communication, it will promote the rapid spread of new ideas and new products on social networks. [0003] To utilize social networks for viral marketing, Kempe et al. [1] proposed the influence maximization problem for the first time: select an initial user set (seed set) of size k, so that under a given propagation model, the number of ultimately affected users is the largest. liter...

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): G06F17/30G06Q50/00
CPCG06Q50/01G06F16/9535
Inventor 刘勇郭龙江王楠李金宝
Owner 江苏派智信息科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com