Author: Kshirsagar, Meghana; Carbonell, Jaime; Klein-Seetharaman, Judith
Title: Multitask learning for host–pathogen protein interactions Document date: 2013_7_1
ID: sdgt2ms5_75
Snippet: We found that the feature weights vary greatly across models-the cosine similarity ranges between 0.1 and 0.13. We also analyzed which features had the highest absolute weight. We found that the node-degree feature (computed using the human PPI graph) has a high positive weight across all tasks. Gene expression features have large negative weights across all tasks. In general, the GO and protein sequence-based n-gram features have different weigh.....
Document: We found that the feature weights vary greatly across models-the cosine similarity ranges between 0.1 and 0.13. We also analyzed which features had the highest absolute weight. We found that the node-degree feature (computed using the human PPI graph) has a high positive weight across all tasks. Gene expression features have large negative weights across all tasks. In general, the GO and protein sequence-based n-gram features have different weights across tasks.
Search related documents:
Co phrase search for related documents- cosine similarity and gene expression: 1, 2, 3
- different weight and gene expression: 1
- different weight and negative weight: 1, 2, 3, 4
- gene expression and greatly vary: 1
- gene expression and human PPI graph: 1
- gene expression and human PPI graph compute: 1
- gene expression and negative weight: 1
- gene expression and node degree: 1
- human PPI graph and node degree: 1, 2
- human PPI graph compute and node degree: 1
- large negative weight and negative weight: 1
Co phrase search for related documents, hyperlinks ordered by date