On measuring confidence levels using multiple views of feature set for useful unlabeled data selection

Multi-view based unlabeled data selection using feature transformation methods for semiboost learning

“…In this section, the SemiBoost [23] algorithm and the related criteria, which are closely related to the present paper, are briefly overviewed in order to make it complete. The detailed description can also be found in the literature [19,20,23].…”

“…Then, a few examples with higher confidence levels are selected to retrain the ensemble classifier together with L. However, it is not guaranteed that adding the selected data to the training data will lead to a situation in which the classification performance can be improved [35]. Therefore, various approaches have been proposed in the literature for selecting a small amount of useful unlabeled data (U s ) from U : these include the self-training [25,30,40] and co-training [3,12,18] approaches, confidence-based approaches [19,20,23], density/distance-based approaches [8,27,28], and other approaches used in active learning (AL) algorithms [7,11,33].…”

“…Motivated by this, in [20], in order to select useful U s efficiently, the confidence levels of all x ∈ U examples were measured using multiple views extracted from the feature set of the data. In the multi-view based strategy, a small amount of useful U s are selected as follows: first, multiple views of the data are derived independently through a decomposition method of the feature set of the data; second, each of the views is used to measure the confidence levels of the data; third, all the confidence levels measured from the views are used as a weighted average to derive a target confidence.…”

“…The main goal of this paper is to demonstrate that using a transformationbased selection (TBS) method when selecting U s from U , instead of the data decomposition-based selection (DBS) method (which has been developed in [20]), leads to an improvement in terms of classification accuracies. In DBS, the feature set of the data is independently decomposed into multiple views.…”

SIP-FS: a novel feature selection for data representation

“…Different from a scalar feature, feature types, which can be scalars, vectors, or matrices, are highly diverse in dimension and expression. However, existing methods simply ensemble the selection of each feature type [13] or concatenate all features types into a single vector [14]. These methods ignore the relation between different feature types.…”