Constraint Selection-Based Semi-supervised Feature Selection

Abstract

Dimensionality reduction is a significant task when dealing with high-dimensional data, this reduction can be done by feature selection, which means to select the most appropriate features for data analysis. It is a recent addressed challenge in feature selection research when handling small-labeled with largeunlabeled data sampled from the same population. The supervision information may be used in the form of pairwise constraints; these constraints have practically proven to have very positive effects on the learning performance. Nevertheless, selected constraints sets may have significant results (positive or negative) on learning performance. In this paper, we present a novel feature selection approach based on an efficient selection of pairwise constraints. This aims to grasp the most coherent constraints extracted from labeled party of data. We then evaluate the relevance of a feature according to its 'efficient' locality preserving and 'chosen' constraints preserving ability. Finally, experimental results will be provided for validating our proposal in comparison with other known feature selection methods.

Soft-constrained Laplacian score for semi-supervised multi-label feature selection

“…More recently, we proposed in [2] a constrained Laplacian score (CLS) for semi-supervised feature selection by a more developed combination between the Laplacian score [17] and the constraint score [19]. In the same way, we improved CLS in [18] by another method (CSFS), which exploits a constraint selection procedure during the feature selection process. Our recent work can be found in [3] with a theoretical analysis of CSFS and a new graph-based procedure to reduce redundancy.…”

A Novel Approach for Semi Supervised Document Clustering with Constraint Score based Feature Supervision

“…It proposed a framework for feature selection based on constraint selection for semi-supervised dimensionality reduction. A new score function was developed to evaluate the relevance of features based on both, the locally geometrical structure of unlabeled data and the constraints preserving ability of labeled data [5].Chen, et al proposed an effective Fuzzy Frequent Itemset-based Document Clustering (F 2 IDC). This approach combines the fuzzy association rule mining with the background knowledge embedded in WordNet, which improve the quality of document clustering [6].…”

Selecting feature subset with sparsity and low redundancy for unsupervised learning

“…In recent years, many research efforts have been devoted to achieving this target. According to the availability of label information, these methods can be roughly grouped into three categories, i.e., supervised feature selection [6], semi-supervised feature selection [7] and unsupervised feature selection [8][9][10]. Typically, supervised and semi-supervised methods both require the label information to varying degree for feature selection [9].…”