CLASSIFICATION OF SAMPLE DATA

摘要

A classification system (100) for classification of bio molecular data is provided. An input of the system receives a plurality of features (102) of a sample to be classified and a plurality of respective error estimates (104). A statistical module (106) associates probability density functions (108) with the features, wherein variances of the probability density functions depend on the error estimates. A replication module (110) produces a plurality of perturbed replicas (112) of the sample, wherein the features are randomly perturbed according to the corresponding respective probability density functions. A classifier (114) classifies the perturbed replicas based on the perturbed features. An analyzer (118) classifies the sample to be classified based on a statistical analysis of the classified replicas (116) to obtain a sample classification (120).