Class imbalance methods for translation initiation site recognition

摘要: Translation initiation sites (TIS) recognition is one of the first steps in gene structure prediction, and common components any system. Many methods have been described literature to identify TIS transcripts such as mRNA, EST cDNA sequences. However, DNA sequences a far more challenging task, so for achieve poor results Most approach this problem taking into account its biological features. In work we try different view, considering classification from purely machine learning perspective.From point view learning, class imbalance problem. Thus, paper angle, apply that developed deal with datasets.Results show an advantage respect same applied without nature The are also able improve obtained best method literature, which based on looking next in-frame stop codon putative must be predicted.