The TF-IDF and Neural Networks Approach for Translation Initiation Site Prediction

The precise prediction of translation initiation site is an important task for the analysis of genomic sequence. This study aims to increase the accuracy for the prediction of translation initiation site using a TF-IDF-NN-TIS model (TF-IDF and neural networks approach for translation initiation site prediction). This study creates feature using 1-gram and 2-gram techniques for both upstream and downstream. Determining feature value uses TF-IDF approach and feature selection by correlation-based feature selection method. Evaluation prediction results use 10-fold cross validation. This study performed experiments on three different datasets that are Vertebrate, Arabidopsis thaliana, and TIS+50. The results of the study indicate that the proposed model gives highest accuracy with less processing time.