Table 2.
AUC score performances for different variations of AttentiveChrome and baselines
| Baselines | AttentiveChrome Variations | ||||||
|---|---|---|---|---|---|---|---|
|
| |||||||
| Model | DeepChrome (CNN) [29] |
LSTM | CNN- Attn |
CNN- α, β |
LSTM- Attn |
LSTM- α |
LSTM- α, β |
|
| |||||||
| Mean | 0.8008 | 0.8052 | 0.7622 | 0.7936 | 0.8100 | 0.8133 | 0.8115 |
| Median | 0.8009 | 0.8036 | 0.7617 | 0.7914 | 0.8118 | 0.8143 | 0.8123 |
| Max | 0.9225 | 0.9185 | 0.8707 | 0.9059 | 0.9155 | 0.9218 | 0.9177 |
| Min | 0.6854 | 0.7073 | 0.6469 | 0.7001 | 0.7237 | 0.7250 | 0.7215 |
|
| |||||||
| Improvement over DeepChrome [29] (out of 56 cell types) | 36 | 0 | 16 | 49 | 50 | 49 | |