Transductive Multiview Modeling With Interpretable Rules, Matrix Factorization, and Cooperative Learning
ABSTARCT :
Multiview fuzzy systems aim to deal with fuzzy modeling in multiview scenarios effectively and to obtain the interpretable model through multiview learning. However, current studies of multiview fuzzy systems still face several challenges, one of which is how to achieve efficient collaboration between multiple views when there are few labeled data. To address this challenge, this article explores a novel transductive multiview fuzzy modeling method. The dependency on labeled data is reduced by integrating transductive learning into the fuzzy model to simultaneously learn both the model and the labels using a novel learning criterion. Matrix factorization is incorporated to further improve the performance of the fuzzy model. In addition, collaborative learning between multiple views is used to enhance the robustness of the model. The experimental results indicate that the proposed method is highly competitive with other multiview learning methods.
EXISTING SYSTEM :
? We focus on the challenge that exists in multi-view clustering and propose a potential solution that enhances the clustering performance.
? Among the existing techniques used for constructing the shared subspace for multiple views, non-negative matrix factorization has attracted extensive attention.
? Different from existing multi-view clustering methods, a new shared hidden space learning method is proposed in the present study to exploit the relationship between different views along with the optimization of the clustering centers.
? Many existing multi-view clustering methods such as Co-FKM, Co-FCM and WV-COFCM, the proposed HSS-MVFC method will continue to treat FCM as the basic framework.
DISADVANTAGE :
? In this paper, we approach this problem by proposing a novel normalization strategy and following the principle that factors representing clustering structures learnt from multiple views should be regularized toward a common consensus.
? We present the proposed joint matrix factorization formulation for multi-view clustering and effective iterative update rules to solve the optimization problem.
? In light of this challenge, we design a novel normalization procedure, which can successfully solve the problem.
? However, due to non-sparseness of the intermediate matrices for eigendecomposition, the computational complexity becomes an issue, which can be observed.
PROPOSED SYSTEM :
• The proposed method uses maximum entropy strategy to control the weights of different views while learning the shared hidden space.
• In different application fields, many multi-view algorithms have been proposed to make full use of multi-view data, such as multiview cooperation learning in clustering.
• In order to deal with large-scale data clustering problems, a new robust large-scale multi-view clustering method was proposed to integrate multiple representations of large scale data.
• Some researchers have proposed multi-view clustering ensemble learning that combines different ensemble techniques for multi-view clustering.
• The framework of FCM is the basic framework of the proposed method and that of NMF is used to construct the shared hidden view.
ADVANTAGE :
? Two metrics, the accuracy (AC) and the normalized mutual information (NMI) are used to measure the clustering performance.
? Observing that these multiple representations often provide compatible and complementary information, it becomes natural for one to integrate them together to obtain better performance rather than relying on a single view.
? NMF has become a popular technique for data clustering, and it is reported to achieve competitive performance compared with most of the state-of-the-art unsupervised algorithms.
? Therefore, in terms of matrix factorization, we require coefficient matrices learnt from different views to be softly regularized towards a common consensus.
|