COMPUTATIONAL METHOD FOR DISCOVERING PATTERNS IN DATA SETS

摘要

Automatic discovery of qualitative and quantitative patterns inherent in datasets isaccomplished by use of a muffed framework which employs adjusted residualanalysis instatistics to test the significance of the pattern candidates generated fromdata sets. Thisframework consists of a search engine for different order patterns, amechanism to avoidexhaustive search by eliminating impossible pattern candidates, an attributedhypergraph(AHG) based knowledge representation language and an inference engine whichmeasuresthe weight of evidence of each pattern for classification and prediction. If apattern candi-date passes the statistical significance test of adjusted residual, it isregarded as a patternand represented by an attributed hyper edge in AHG. In the task ofclassification and/orprediction, the weights of evidence are calculated and compared to draw theconclusion.