Table 1.
Performance evaluation of identifying significant time intervals from simulated features
|
σ = 5 |
σ = 1 |
||||
|---|---|---|---|---|---|
| Sensitivity | Specificity | Sensitivity | Specificity | ||
| Pattern_1 (early change) |
OmicsLonDA (SSANOVA) | 0.976 | 0.998 | 0.982 | 0.993 |
| OmicsLonDA (GAMM) | 0.945 | 0.991 | 0.965 | 0.990 | |
| MetaLonDA | 0.993 | 0.833 | 0.992 | 0.881 | |
| splinectomeR | 1 | 0.698 | 1 | 0.735 | |
| Pattern_2 (middle change) |
OmicsLonDA (SSANOVA) | 0.924 | 0.999 | 0.952 | 0.998 |
| OmicsLonDA (GAMM) | 0.948 | 0.995 | 0.963 | 0.997 | |
| MetaLonDA | 0.991 | 0.818 | 0.991 | 0.818 | |
| splinectomeR | 0.999 | 0.682 | 1 | 0.699 | |
| Pattern_3 (late change) |
OmicsLonDA (SSANOVA) | 0.906 | 0.999 | 0.921 | 0.999 |
| OmicsLonDA (GAMM) | 0.934 | 0.995 | 0.957 | 0.998 | |
| MetaLonDA | 0.984 | 0.805 | 0.991 | 0.876 | |
| splinectomeR | 0.965 | 0.804 | 1 | 0.772 | |
| Pattern_4 (very late change) |
OmicsLonDA (SSANOVA) | 0.870 | 0.998 | 0.902 | 0.999 |
| OmicsLonDA (GAMM) | 0.725 | 0.999 | 0.774 | 0.999 | |
| MetaLonDA | 0.926 | 0.795 | 0.932 | 0.829 | |
| splinectomeR | 0.995 | 0.729 | 1 | 0.765 | |
| Pattern_5 (no change) |
OmicsLonDA (SSANOVA) | — | 0.982 | — | 0.983 |
| OmicsLonDA (GAMM) | — | 0.989 | — | 0.994 | |
| MetaLonDA | — | 0.972 | — | 0.981 | |
| splinectomeR | — | 0.959 | — | 0.962 | |