Skip to main content

Table 10 Comparison of accuracy using binary methylation classification.

From: Profile analysis and prediction of tissue-specific CpG island methylation classes

Methylation classification Dataset Methods Acc[%] CC
binary HEP1 EpiGRAPH – SVM linear kernel* 84.90 0.657
binary HEP1 EpiGRAPH – Decision tree C4.5* 75.80 0.462
binary HEP1 Matlab – Decision tree** 90.08 0.743
binary EpiGRAPH 2 EpiGRAPH – SVM linear kernel* 85.20 0.658
binary EpiGRAPH 2 EpiGRAPH – Decision tree C4.5* 78.60 0.524
binary EpiGRAPH 2 Matlab – Decision tree** 91.67 0.775
four classes HEP3 Matlab – Decision tree** 89.39 0.707
  1. *Validation was performed using 10 repetitions of 10 fold cross-validation
  2. ** Validation was performed using 10 fold cross-validation
  3. 1HEP CGI data (using our attributes and binary methylation classes)
  4. 2EpiGRAPH methylation data (using the default EpiGRAPH sequence attributes and binary methylation classes)
  5. 3HEP CGI data (using our attributes and four methylation classes)
  6. The average accuracy (Acc) and correlation coefficient (CC) were used to measure fitness.