Research on User Influence Weighted Scoring Algorithm Incorporating Incentive Mechanism

EasyChair Preprint 9525

Abstract

Overall product ratings are an important basis for users when shopping online or using online services. However, some sellers and web service providers put a large amount of false rating data into the rating system to improve their own rankings, which seriously damages the interests of users. In this paper, two methods are used to reduce the impact of false ratings on overall ratings. First, a user influence weighted scoring algorithm is proposed to analyze user behavior and build a user influence model. The influence of different users on the rating is considered when calculating the overall rating to improve the accuracy of the project's overall rating. Secondly, a blockchain-based rating incentive mechanism is designed to correlate users' rating behavior with their interests, effectively constraining their rating behavior and making them consciously and proactively provide more authentic ratings. Simulations comparing the proposed algorithm with the rating algorithms used on Douban and IMDB show that the algorithm performs best in terms of resistance to interference. The experimental results also show that the rating incentive mechanism can reward high-impact users and punish low-impact malicious users, and can effectively defend against malicious users.

Keyphrases: Blockchain, User influence model, Weighted scoring algorithm, incentive mechanism

@booklet{EasyChair:9525,
  author    = {Jingya Xu and Lina Ge and Wei Zhou and Liang Yan and Zheng Hu},
  title     = {Research on User Influence Weighted Scoring Algorithm Incorporating Incentive Mechanism},
  howpublished = {EasyChair Preprint 9525},
  year      = {EasyChair, 2023}}