基于协同过滤算法的自动化隐式评分音乐双重推荐系统

An automated implicit scoring music dual recommendation system based on collaborative filtering algorithm

摘要:

根据传统系统采用的显示评分机制受到外界干扰影响，使推荐结果精准度低的问题，提出了基于协同过滤算法的自动化隐式评分音乐双重推荐系统。在异构普适环境推荐框架下，对系统总体结构进行设计。其中硬件部分采用四元件组成方式，使用W900710型号芯片作为播放器核心板，并将隐式评分提取器与推荐引擎结合起来，可避免噪声干扰。而软件部分设计场景模拟衰减现象，采用协同过滤算法描述衰减过程，根据描述结果，设立双重推荐机制来实现抗人为影响的音乐双重推荐系统。由实验结果可知，采用协同过滤算法设计的系统最高推荐结果精准度可达到90%，对于大规模音乐数据推荐具有良好可扩展性。

Abstract:

According to the influence of the display scoring mechanism adopted by the traditional system, which is influenced by the external interference and the precision of the recommended results is low, a dual recommendation system based on collaborative filtering algorithm is proposed. Under the recommendation framework of heterogeneous ubiquitous environment, the overall structure of the system is designed. The hardware part is composed of four components, using W900710 model chip as the core of the player, and combining the implicit score extractor with the recommendation engine, the noise interference can be avoided. The software part design scene simulation attenuation phenomenon, use collaborative filtering algorithm to describe the attenuation process, according to the description results, set up a double recommendation mechanism to achieve the anti human impact of the music dual recommendation system. The experimental results show that the maximum precision of the proposed system can reach 90%, and it has good scalability for large-scale music data recommendation.