Research Article
A Novel Two-Stage Spectrum-Based Approach for Dimensionality Reduction: A Case Study on the Recognition of Handwritten Numerals
Algorithm 2
Two-stage dimensionality reduction procedure.
Extract most-used features (in literature) from input dataset, and Create initial features set Initial_; | Number of features in initial features set Initial_; | first reduced version of feature vector null; | Stage 1: | Choosing threshold , by investigating 1D_MM spectrum diagrams. | for ( : ) | { | Compute the coordinate of all 1D_SD spectrum lines corresponding to feature ; | for ( : number of classes) | { | If (overlapping of spectrum line of class with all the rest spectrum lines has the value less than threshold ) then | { | Insert feature to ; | goto ; | } | } | : continue; | } | Number of features in first reduced version of features vector ; | final reduced versions of feature vectors null; | Stage 2: | Choosing threshold , by investigating 2D_MM spectrum diagrams. | for ( : ) | for ( : ) | { | Compute the coordinate of all 2D_SD spectrum ellipses corresponding to features pair (, ); | for ( : number of classes) | { | if (overlapping of spectrum ellipses of class with all the rest spectrum ellipses has the value less than | threshold ) then | { | Insert feature pair and to ; | goto ; | } | } | } | : continue; | } | delete the repetitive features from ; |
|