Improving feature selection algorithms using normalised feature histograms
This addresses instability issues in feature selection for bioinformatics and image analysis applications, though it appears incremental as it builds on existing methods.
The paper tackled the instability problem in conventional feature selection methods by proposing a method that builds histograms of stable features from random training subsets and ranks them using classifier-based cross-validation. Results showed considerable improvement over benchmarks on four microarray and three image datasets using three feature selection criteria with a naive Bayes classifier.
The proposed feature selection method builds a histogram of the most stable features from random subsets of a training set and ranks the features based on a classifier based cross-validation. This approach reduces the instability of features obtained by conventional feature selection methods that occur with variation in training data and selection criteria. Classification results on four microarray and three image datasets using three major feature selection criteria and a naive Bayes classifier show considerable improvement over benchmark results.