FEATURE VECTOR-BASED METHOD FOR REMOVING REDUNDANCY IN A TRAINING DATASET

摘要

PURPOSE: A repetition removal method of training data based on a feature vector is provided to include a feature of protein and an RNA(RiboNucleic Acid) sequence in the feature vector as a composition element, thereby effectively predicting an RNA-combination amino acid existed in a protein sequence. CONSTITUTION: A protein-RNA combination part determines an RNA-combination amino acid which interacts with RNA using a hydrogen bond(S10). A tendency property of an amino acid triplet is calculated(S20). Various features of protein and RNA sequences are coded in a feature vector in order to predict the RNA-combination amino acid(S30). A training data set, which removes the unnecessary repetition of data, is constructed based on the coded feature vector(S40). [Reference numerals] (S10) Step of determining a protein-RNA combination part; (S20) Step of calculating a tendency property of an amino acid triplet; (S30) Step of coding in a feature vector; (S40) Step of constructing a training data set;