Supervised Fractional-Order Embedding Geometrical Multi-View CCA (SFGMCCA) for Multiple Feature Integration

Abstract

Techniques for integrating different types of multiple features effectively have been actively studied in recent years. Multiset canonical correlation analysis (MCCA), which maximizes the sum of pairwise correlations of interview (i.e., between different features), is one of the powerful methods for integrating different types of multiple features, and various MCCA-based methods have been proposed. This work focuses on a supervised MCCA variant in order to construct a novel effective feature integration framework. In this paper, we newly propose supervised fractional-order embedding geometrical multiview CCA (SFGMCCA). This method constructs not only the correlation structure but also two types of geometrical structures of intra-view (i.e., within each feature) and interview simultaneously, thereby realizing more precise feature integration. This method also supports the integration of small sample and high-dimensional data by using the fractional-order technique. We conducted experiments using four types of image datasets, i.e., MNIST, COIL-20, ETH-80 and CIFAR-10. Furthermore, we also performed an fMRI dataset containing brain signals to verify the robustness. As a result, it was confirmed that accuracy improvements using SFGMCCA were statistically significant at the significance level of 0.05 compared to those using conventional representative MCCA-based methods. INDEX TERMS Feature integration, multi-view, canonical correlation analysis, fractional-order technique, geometrical structure.

Trial Analysis of the Relationship between Taste and Biological Information Obtained While Eating Strawberries for Sensory Evaluation

“…As a solution to this problem, multivariate analysis models that introduce label dequantization have been proposed [ 48 , 49 ], and their use will be considered in future work. In addition, while general CCA is based on two types of data, methods based on multi-view CCA are effective when various types of data are used [ 50 , 51 ], as in our experiment in this paper. Capturing more complex correlations will become feasible by introducing the above methods.…”

Detection of Important Scenes in Baseball Videos via Bidirectional Time Lag Aware Deep Multiset Canonical Correlation Analysis

“…(1), we obtain the general eigenvalue problem of Eq. (7) [32], [33]. BiTl-dMCCA is an extended version of dMCCA that solves the problem.…”

Multi-Modal Label Dequantized Gaussian Process Latent Variable Model for Ordinal Label Estimation

“…CCA can represent a lower-dimensional common latent space in This work was partly supported by the MIC/SCOPE #181601001. such a way that a pair correlation of the multi-view data is maximized. In particular, CCA which can be applied to three or more kinds of modalities is called multi-view CCA (MVCCA) [5][6][7][8]. Furthermore, to realize more accurate representation of the lowerdimensional common latent space, some advanced MVCCA-based approaches have been studied [9,10].…”

“…Besides, a method utilizing label information based on mG-PLVM has been proposed [15], and this method considers intra-and inter-class correlations in the same manner as linear discriminant analysis (LDA) [16]. Incidentally, a recent study [8] have revealed that the direct use of label information as one modality can estimate the more effective common latent space than correlation coefficientbased approaches. Thus, by introducing label information as one of the modalities into mGPLVM, it is expected that better representation of the common latent space can be obtained.…”