A personalized music recommendation system and its implementation method based on multidimensional time series analysis

Abstract

The invention relates to a system for recommending individual music based on multi-dimensional time series analysis and an achieving method of the system. The system comprises a system interaction interface, a candidate song database, a user behavior database and a time series recommendation engine. According to the system and the achieving method, the diversity of song classifying standards and the time series of the listening behavior of a user are taken into consideration fully, the listening behavior of the user is modeled into the multi-dimensional time series based on the fact that probability distribution of a plurality of implied themes is taken into consideration in the process of song modeling, the behavior habits of the user are explored through the method of multi-dimensional time series analyzing, finally appropriate songs are recommended to the user from the candidate song database, and the accuracy of recommendation is improved.

Description

technical field [0001] The invention relates to a personalized music recommendation system based on multidimensional time series analysis and its realization method. Background technique [0002] The vigorous development of digital music and Internet technology has greatly accelerated the appearance and dissemination of songs, while the emergence of cloud services and smart phones has provided people with a more convenient way to collect or listen to a large number of music works. However, the massive number of songs has caused a serious problem of information overload, making it impossible for people to quickly obtain songs that meet their preferences. In order to deal with this problem, the music recommendation system came into being, and has become a very potential method to solve the problem of information overload in the music field. [0003] The music recommendation system is essentially an information filtering system, which helps users filter out unnecessary informa...

AI Technical Summary

Problems solved by technology

This is mainly manifested in the following points: 1. There are many classification methods and standards for music, and these standards are often not exclusive, which makes it difficult to use a set of unified and accurate features to describe songs; 2. Traditional recommendation algorithms seldom consider Songs are short in time, low in consumption cost, and easy to form sequences; 3. Songs have strong emotional color, and users' consumption of songs is more sensitive to the context in which they are located, and the sequence formed by users listening to songs can just reflect this. context

Images