IEEE Transactions on Cybernetics · 2013 · 217 citations · 36 references
ComputingComputer ScienceFeature EngineeringMachine LearningStatistical LearningDimensionality ReductionOptimization For Machine LearningOutlier DetectionOutlier HandlingOutlier TreatmentPattern RecognitionIterative LearningLinear Discriminant AnalysisRecognition ProblemsSubspace Learning
Fisher linear discriminant analysis (LDA) is a classical subspace learning technique for pattern recognition that relies on an L2‑norm criterion, making it vulnerable to outliers. This paper introduces LDA‑L1, a method that maximizes the ratio of between‑class to within‑class dispersion using an L1‑norm to enhance robustness. LDA‑L1 is solved by an iterative algorithm that converges to a locally maximal point, is easy to implement, and provides robustness against outliers. The method eliminates the small‑sample size and rank‑deficiency problems of conventional LDA and demonstrates superior image‑recognition performance in experimental tests.
Fisher linear discriminant analysis (LDA) is a classical subspace learning technique of extracting discriminative features for pattern recognition problems. The formulation of the Fisher criterion is based on the L2-norm, which makes LDA prone to being affected by the presence of outliers. In this paper, we propose a new method, termed LDA-L1, by maximizing the ratio of the between-class dispersion to the within-class dispersion using the L1-norm rather than the L2-norm. LDA-L1 is robust to outliers, and is solved by an iterative algorithm proposed. The algorithm is easy to be implemented and is theoretically shown to arrive at a locally maximal point. LDA-L1 does not suffer from the problems of small sample size and rank limit as existed in the conventional LDA. Experiment results of image recognition confirm the effectiveness of the proposed method.
36
UCI Machine Learning Repository
Arthur Asuncion · Medical Entomology and Zoology · 2007 · 24.4K citations
THE USE OF MULTIPLE MEASUREMENTS IN TAXONOMIC PROBLEMS
Ronald Aylmer Fisher · Annals of Eugenics · 1936 · 14.7K citations · Full text
Matthew Turk, Alex Pentland · Journal of Cognitive Neuroscience · 1991 · 13.8K citations · Full text
Statistical Comparisons of Classifiers over Multiple Data Sets
Janez Demšar · 2006 · 11.2K citations