Supervised multiview learning based on simultaneous learning of multiview intact and single view classifier

Abstract

Multiview learning problem refers to the problem of learning a classifier from multiple view data. In this data set, each data points is presented by multiple different views. In this paper, we propose a novel method for this problem. This method is based on two assumptions. The first assumption is that each data point has an intact feature vector, and each view is obtained by a linear transformation from the intact vector. The second assumption is that the intact vectors are discriminative, and in the intact space, we have a linear classifier to separate the positive class from the negative class. We define an intact vector for each data point, and a view-conditional transformation matrix for each view, and propose to reconstruct the multiple view feature vectors by the product of the corresponding intact vectors and transformation matrices. Moreover, we also propose a linear classifier in the intact space, and learn it jointly with the intact vectors. The learning problem is modeled by a minimization problem, and the objective function is composed of a Cauchy error estimator-based view-conditional reconstruction term over all data points and views, and a classification error term measured by hinge loss over all the intact vectors of all the data points. Some regularization terms are also imposed to different variables in the objective function. The minimization problem is solve by an iterative algorithm using alternate optimization strategy and gradient descent algorithm. The proposed algorithm shows it advantage in the compression to other multiview learning algorithms on benchmark data sets.

Semi-supervised Multi-view Manifold Discriminant Intact Space Lear

“…In this section, we compare the proposed SM 2 DIS approach with several state-of-the-art methods on the above four databases, including three types of methods: semi-supervised multi-view feature learning methods: SULF [25], MCL [26], PSLF [27], AMVNMF [28]; unsupervised multi-view feature learning method MISL [13]; supervised multi-view feature learning method MISC [18].…”

“…Some researchers utilize the label information of multi-view data to extract effect discriminant information [14][15][16][17][18][19][20][21]. For many real-world applications such as image annotation, gene function prediction, and insider threat detection, hierarchical multi-latent space (HiMLS) [16] learns a hierarchical multi-latent space to jointly model the triple types of heterogeneity, i.e., task heterogeneity, view heterogeneity, and label heterogeneity of the data.…”

“…By maximally preserving the discriminative information in the learned latent subspace, DMINTIR losslessly recovers the original feature space. Based on the assumption that there exists one single discriminative intact feature representation for different views of the data point, multiview intact and single-view classifier learning (MISC) [18] tries to recover the discriminative intact feature space for the data points and learns a classifier in the intact space with the help of class labels of the data. However, some favorable information in unlabeled samples cannot be effectively exploited.…”

Individuality- and Commonality-Based Multiview Multilabel Learning

“…For instance, an image can be simultaneously annotated with several labels, such as sea, sky, and seagull; a web page could be tagged with multiple topics given as labels, such as economics, culture, sports and politics. On the other hand, the majority of existing studies are supervised approaches that require a large number of labeled samples [5], [6]. In practice, nevertheless, it is rather difficult and expensive to collect labeled samples, while unlabeled samples are easy to accumulate.…”

Salary Prediction in the IT Job Market with Few High-Dimensional Samples: A Spanish Case Study

“…These models are often used as the baseline in several contexts and sometimes are sufficient for a highly accurate prediction (see Refs. 27,28). In order to quantify the performance of a linear model, we consider the coefficient of determination (R 2 ), which provides an assessment of the variability of the output any linear model is able to capture and depends on the variance of the data and the sum of the squared errors of the model.…”