BP神经网络算法在音乐流行趋势预测中的应用研究

发布时间：2018-02-28 14:09

本文关键词： 神经网络指数平滑法 ARIMA模型音乐预测　出处：《西南交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：音乐的流行趋势可以根据当前的流行艺人表现出来,因此对音乐流行趋势的预测也就是对哪些音乐艺人即将成为未来一段时间内的流行艺人的预测。而判断某个艺人是否是流行艺人则可以根据该艺人最近一段时间里的音乐试听量来判断。通过统计分析用户对音乐的操作(试听、下载、收藏)记录,预测出艺人在下一阶段内的音乐试听量,从而可以判断出哪些艺人在未来一段时间内音乐试听量最高,这些艺人即代表着未来一段时间内的音乐流行趋势。本文通过统计分析电子音乐平台产生的用户试听、下载、收藏歌曲的行为记录,结合二次指数平滑法、自回归移动平均模型以及BP神经网络模型对艺人歌曲试听量进行了预测,同时设计并实现了基于BP神经网络算法的音乐流行趋势预测系统。本文的主要研究工作如下:1.通过阅读大量国内外文献,研究了国内外音乐试听量预测的研究现状、神经网络算法研究的现状、该算法的特点以及在多个应用领域中的使用情况。重点研究了BP神经网络算法的应用,并对电子音乐平台上产生的基础数据进行统计分析,寻找影响艺人音乐试听量的主要因素,最后使用BP神经网络算法对艺人在接下来一个月内每天的音乐试听总量进行了预测。2.在使用BP神经网络算法对艺人音乐试听量进行预测的同时,使用了二次指数平滑法、自回归移动平均模型对艺人音乐试听量进行预测,最后对比三种预测方法的预测结果。3.设计并实现了基于BP神经网络算法的音乐流行趋势预测系统。该系统是通过J2EE平台,结合web开发技术、数据库技术、数据挖掘技术进行开发的。可以让不懂二次指数平滑法、自回归移动平均模型、BP神经网络算法等预测算法的工作人员也能通过操此系统预测艺人在下一阶段中的音乐试听量,从而判断出哪些艺人即将代表下一阶段的音乐流行趋势。最后总结了本文的研究内容,并对下一步的工作作出展望。
[Abstract]:The pop trend of music can be expressed according to the current pop artists. Therefore, the prediction of pop trends is a prediction of which musical artists will become pop artists for some time to come. And judging whether an artist is a pop artist can be based on the latest period of time. Through the statistical analysis of the user's operation of the music (listen to, listen to, Download, collect) records to predict the amount of music auditions that artists will be listening to in the next stage, so as to determine which artists will have the highest amount of music auditions over the next period of time. These artists represent the trend of music popularity for a period of time in the future. This paper, through statistical analysis of the users' listening, downloading and collecting songs generated by the electronic music platform, combines the quadratic exponential smoothing method. The autoregressive moving average model and BP neural network model were used to predict the audition quantity of artist songs. The main research work of this paper is as follows: 1. By reading a large number of domestic and foreign literature, the research status of music audition prediction at home and abroad is studied. The present situation of neural network algorithm research, the characteristics of the algorithm and its application in many application fields are discussed. The application of BP neural network algorithm is studied, and the basic data generated on the electronic music platform are analyzed statistically. Look for the main factors that affect the amount of music auditions by artists, Finally, BP neural network algorithm is used to predict the total amount of music auditions of artists in the next month. 2. While using BP neural network algorithm to predict the amount of music auditions of artists, the quadratic exponential smoothing method is used. The Auto-regressive moving average model is used to predict the audition quantity of entertainers. Finally, the prediction results of three prediction methods are compared. 3. A music trend prediction system based on BP neural network algorithm is designed and implemented. The system is based on J2EE platform. Combined with web development technology, database technology, data mining technology to develop. Can not understand the quadratic exponential smoothing method, Staff members of prediction algorithms such as the autoregressive moving average model and BP neural network algorithm can also use this system to predict the music auditions of artists in the next stage. In order to judge which artists will represent the next stage of music trends. Finally, this paper summarizes the content of the study, and makes a prospect for the next work.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.52;TP183

【参考文献】