Selected article for: "training set and validation training set"

Author: Jakub M Bartoszewicz; Anja Seidel; Bernhard Y Renard
Title: Interpretable detection of novel human viruses from genome sequencing data
  • Document date: 2020_1_30
  • ID: ac00tai9_15
    Snippet: For the positive class, we generated a training set containing 80% of the genomes, and validation and test sets with 10% of the genomes each. Importantly, the nested structure was kept also during the training-validation-test split: for example, the species assigned to the smallest test set ("Chordata") were also present in all the bigger test sets. The same applied to other taxonomic levels, as well as the training and validation sets wherever a.....
    Document: For the positive class, we generated a training set containing 80% of the genomes, and validation and test sets with 10% of the genomes each. Importantly, the nested structure was kept also during the training-validation-test split: for example, the species assigned to the smallest test set ("Chordata") were also present in all the bigger test sets. The same applied to other taxonomic levels, as well as the training and validation sets wherever applicable.

    Search related documents:
    Co phrase search for related documents