Table 5.
The best prediction results of E. coli promoters obtained by different methods (fragments length is 80 bp)**
| Methods | Results (%) |
||
|---|---|---|---|
| Sensitivity TP/(TP+FN) | Specificity TN/(TN+FP) | Precision TP/(TP+FP) | |
| Negative samples: Coding segments | |||
| IPMD (8) | 84.9 | 91.4 | – |
| Sequence Alignment Kernel+SVM (20) | 82 | – | 84 |
| The proposed method | 96.32 | 95.79 | 95.81 |
| Negative samples: Intergenic segments | |||
| 3-gram* (4) | 67.75 | 86.10 | – |
| IPMD (8) | 81 | 92.7 | – |
| Sequence Alignment Kernel+SVM (20) | 81 | – | 81 |
| The proposed method | 92.11 | 88.77 | 89.13 |
*The negative sample set contained 709 sequence fragments from the coding region and 709 sequence segments from intergenic portions. Training data set size for E. coli was 1669. The paper did not give more details about the training and testing set.
**The best average accuracies among the algorithms evaluated here were shown in boldface.