一种新型电力大数据协同过滤推荐算法的设计与应用

Design and application of a new type of electric power big data collaborative filtering recommendation algorithm

为了解决传统协同过滤推荐算法未考虑知识关联度而导致推荐结果的准确度较低的问题,本文在协同过滤技术与关联规则挖掘技术的基础之上,提出了一种新型电力大数据协同过滤推荐算法(联合算法AR-Item CF),以电力运行系统日志为载体,深入挖掘知识项行为,按照不同知识项行为的轻重程度来划分为不同的权重,以此来计算得出用户评分,将知识关联度作为影响推荐结果排名的主要因素之一,并将目标之间的相似度与关联度均纳入到影响因素之中,推荐结果既存在着关联性,又具备相似性。为了验证算法的可行性,开展了实验对比,测试指标选用MAE与1RR,实验结果表明:本文所提出的新型电力大数据协同过滤推荐算法能够在很大程度上增强推荐系统的推荐效率与推荐质量,取得较佳的效果。

In order to solve the problem that the traditional collaborative filtering recommendation algorithm does not consider the degree of knowledge association,which leads to the low accuracy of the recommendation result,this paper proposes a new method based on collaborative filtering and association rule mining,a new collaborative filtering recommendation algorithm(joint algorithm AR-Item CF)for large power data is proposed,which takes power operation system log as the carrier to mine knowledge item behavior deeply,according to the degree of different knowledge item behavior,it is divided into different weights to calculate the user score,and the degree of knowledge relevance is one of the main factors that influence the ranking of the recommended results,and the similarity and relevance between the goals are taken into account in the influencing factors,the recommendation results have both relevance and similarity.In order to verify the feasibility of the algorithm,an experimental comparison was carried out,and the test indexes were selected as MAE and IRR,the experimental results show that the proposed collaborative filtering algorithm can enhance the recommendation efficiency and quality of the system to a great extent,and achieve better results.