HYBRID METODE HIDDEN MARKOV MODEL DENGAN SELF ORGANIZATION MAPS UNTUK IDENTIFIKASI PROTEIN CODING REGIONS PADA GEN DNA DARI ARABIDOPSIS THALIANA

Main issues to identify the order DNA in genome eurkaryote are processed through the process computing has not been solved. The method Hidden Markov can overcome Model is applied to these issues, which can be classify, and dis-integrity data structure DNA. This Research suggests the application methods Hybrid well with SOM to address the problems in the region identification genome eurkaryote. The SOM algorithm is applied to overcome the shortage method well especially overfitting case. Data that is used in this research is genome plants with latin name Arabidopsis Thaliana. Accuracy of Hybrid method HMM with SOM will be compared with the HMM method will be evaluated on two levels, the level nucleotide and exon. At a level nucleotide will test in terms of sensitivity(Sn), specificity(Sp) and the drag coefficient correlation(CC), in the same thing at a level exon that will test in terms of sensitivity(Sn) and specificity (Sp). Test result by applying methods HMM in the SnHMM = 94% (level nukleotida), SpHMM = 90% (level nukleotida), CCHMM = 2:9%, SnHMM = 82% (level exon), and SnHMM = 90% (level exon). Different with test result in Hybrid method HMM with SOM shows the SnHybrid = 98% (level nukleotida), SpHybrid = 94% (level nukleotida), CcHybrid = 2:6%, SnHybrid = 88% (level exon), and SnHybrid = 94% (level exon).