Categorization of species based on their microRNAs employing sequence motifs, information-theoretic sequence feature extraction, and k-mers

EURASIP Journal on Advances in Signal Processing

Table 1 Average performance of models trained to classify into Hominidae or one of the listed clades

The table shows a comparison between IT and k-mers + sequence motif features. For the k-mers + motif, the best 100 features were selected based on IG (information gain). For the IT, the whole set of features was used. The training/testing was performed with 80%/20% training-testing split, and the average results of 100-fold MCCVs are presented. The table is sorted according to average model accuracy. Note that for the test Hominidae versus H. sapiens, the H. sapiens examples were removed from Hominidae. Highest performance per statistics is highlighted in gray