SVM: Support Vector Machines

Process Initialization Dialog

Parameter Information


Sample Selection

The sample selection option indicates whether to cluster genes or experiments.

Process Selection

The SVM algorithm works performing two main processes, training and classification. One can elect to perform training only, classification only, or both phases of the SVM classification technique.

The 'Training Only' option results in a set of numerical weights which can stored as a file and used for classification at a later time.

The 'Classification Only' option takes a file input of weights generated from training and results in a binary classification of the elements.

The 'Training and Classification' option provides the ability to use the input set as a training set to produce weights which are immediately applied to perform the classification.

The 'One-out Iterative Validation' iteratively performs an SVM training and classification run. On each iteration one element is moved to the neutral classification and therefore will not impact the SVM training nor the classification of elements. The final classification will not be biased by an initial classification of the element.

** Note that the One-out Iterative Validation will iterate a training and classification run for each element being classified. This may not be practical nor necessary for the classification of large numbers of genes. In the case of genes the impact of one gene in producing a discriminate is relatively small and the time to iteratively run for each gene could be prohibitive for large sets of genes. This feature is more practical to use when classifying experiments.

Hierarchical Clustering

This check box selects whether to perform hierarchical clustering on the elements in each cluster created.