Abstract:Recommending music based on a user's music preference is a way to improve user listening experience. Finding the correlation between the user data (e.g., location, time of the day, music listening history, emotion, etc.) and the music is a challenging task. In this paper, we propose an emotion-aware personalized music recommendation system (EPMRS) to extract the correlation between the user data and the music. To achieve this correlation, we combine the outputs of two approaches: the deep convolutional neural networks (DCNN) approach and the weighted feature extraction (WFE) approach. The DCNN approach is used to extract the latent features from music data (e.g., audio signals and corresponding metadata) for classification. In the WFE approach, we generate the implicit user rating for music to extract the correlation between the user data and the music data. In the WFE approach, we use the term-frequency and inverse document frequency (TF-IDF) approach to generate the implicit user ratings for the music. Later, the EPMRS recommends songs to the user based on calculated implicit user rating for the music. We use the million songs dataset (MSD) to train the EPMRS. For performance comparison, we take the content similarity music recommendation system (CSMRS) as well as the personalized music recommendation system based on electroencephalography feedback (PMRSE) as the baseline systems. Experimental results show that the EPMRS produces better accuracy of music recommendations than the CSMRS and the PMRSE. Moreover, we build the Android and iOS APPs to get realistic data of user experience on the EPMRS. The collected feedback from anonymous users also show that the EPMRS sufficiently reflect their preference on music.

No abstract

In this paper, we investigate an attention function combined with the gated recurrent units (GRUs), named GRUA, to raise the accuracy of the customer preference prediction. The attention function extracts the important product features by using the time-bias parameter and the term frequency-inverse document frequency parameter for recommending products to a customer in the ongoing session. We show that the attention function with the GRUs can learn the customer's intention in the ongoing session more precisely than the existing session-based recommendation (SBR) methods. The experimental results show that the GRUA outperforms two SBR methods: the stacked denoising autoencoders with collaborative filtering (SDAE/CF) and the GRUs with collaborative filtering (GRU/CF) based on the precision and recall evaluation metrics. The data from three publicly available datasets, the Amazon Product Review dataset, the Xing dataset, and the Yoo-Choose Click dataset, are used to evaluate the performance of the GRUA with the SDAE/CF and the GRU/CF. This paper shows that adopting the attention function into the GRUs can dramatically increase the accuracy of the product recommendation in the SBR.

No abstract

No abstract