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_18
Snippet: In each of the evaluated settings, we used a total of 20 million (80%) reads for training, 2.5 million (10%) reads for validation and 2.5 million (10%) paired reads as the heldout test set. Read number per genome was proportional to genome length, keeping the coverage uniform on average. While the original datasets are heavily imbalanced, we generated the same number of negative and positive data points (reads) regardless of the negative class de.....
Document: In each of the evaluated settings, we used a total of 20 million (80%) reads for training, 2.5 million (10%) reads for validation and 2.5 million (10%) paired reads as the heldout test set. Read number per genome was proportional to genome length, keeping the coverage uniform on average. While the original datasets are heavily imbalanced, we generated the same number of negative and positive data points (reads) regardless of the negative class definition used.
Search related documents:
Co phrase search for related documents, hyperlinks ordered by date