Most deep learning-based time series clustering models concentrate on data representation in a separate process from clustering. This leads to that clustering loss cannot guide feature extraction. Moreover, most methods solely analyze data from the temporal domain, disregarding the potential within the frequency domain. To address these challenges, we introduce a novel end-to-end Cross-Domain Contrastive learning model for time series Clustering (CDCC). Firstly, it integrates the clustering process and feature extraction using contrastive constraints at both cluster-level and instance-level. Secondly, the data is encoded simultaneously in both temporal and frequency domains, leveraging contrastive learning to enhance within-domain representation. Thirdly, cross-domain constraints are proposed to align the latent representations and category distribution across domains. With the above strategies, CDCC not only achieves end-to-end output but also effectively integrates frequency domains. Extensive experiments and visualization analysis are conducted on 40 time series datasets from UCR, demonstrating the superior performance of the proposed model.

To analyze and manage data, multiple features are extracted for robust and accurate description. Taking account of the complementary nature of multiview data, it is urgent to find a low-dimensional compact representation by leveraging multiple views and generating a better performance for clustering. In this paper, we present multiview latent representation learning with feature diversity for clustering (MvLRFD). To get rid of noise and maintain the information of each view, matrix factorization is adopted to get the latent representation. For fusing the diverse latent representations, we argue that the final representation is constructed by concatenating the latent representation of each view. With the help of the Hilbert–Schmidt Independence Criterion (HSIC), the diversity of features in the final representation is maximized to exploit the complementary information. A new coregularization strategy is introduced to maximize the distance between the similarity between data of the final representation and the latent one. The partition entropy regularization is adopted to control the uniformity level of the value of the similarity matrix and the weight of each view. To verify the availability of MvLRFD compared to state-of-the-art algorithms, several experiments are conducted on four real datasets.

The task of providing recommendations for new users/items that do not have any historical ratings, i.e., the cold-start challenge, has been a subject of extensive research in recommendation systems. Most existing matrix factorization-based approaches tackle this challenge by establishing a relationship between latent profiles and attributes (side information), neglecting the impact of users' and items’ individual characteristics on profile inference. However, individuality is a crucial factor that can significantly influence user/item profiles independently of attributes. To address this issue, we propose constructing the profiles of users/items using an asymmetric mixture distribution of individuality and attribute. This enables the integration of partially observed attributes and unobservable individuality into the profile mixture for improving profile inference. Based on this profile mixture, we introduce the Asymmetric Mixture Factorization (AMF) model, which infers the user/item profiles from Gaussian distributions associated with both attributes and individuality to achieve superior cold-start recommendations. Additionally, we employ an efficient block coordinate descent algorithm and an expectation-maximization approach for optimizing. Empirical evaluations on two datasets demonstrate the effectiveness of our proposed model in handling two cold-start scenarios and achieving state-of-the-art performance.