Table 2. Misclassification rate of class predication methods.
| Type | IHC groups | Grouping as per NCC output | ||||
|---|---|---|---|---|---|---|
| Basal | HER2+ | Luminal A | Luminal B | Error.rate | ||
| A: Classification error rate: NCC centriod method | ||||||
| Basal | 35 | 31 | 3 | 1 | 0 | 0.11 |
| HER2+ | 25 | 3 | 15 | 3 | 4 | 0.4 |
| Luminal A | 27 | 1 | 2 | 21 | 3 | 0.22 |
| Luminal B | 30 | 5 | 3 | 13 | 9 | 0.7 |
| Type | IHC groups | Grouping as per Random forest output | |||||
|---|---|---|---|---|---|---|---|
| Total | Basal | HER2+ | Luminal A | Luminal B | Error.rate | ||
| B: Classificaon error rate: random forest method | |||||||
| Basal | 35 | 30 | 4 | 1 | 0 | 0.14 | |
| HER2+ | 25 | 2 | 19 | 1 | 3 | 0.24 | |
| Luminal A | 27 | 1 | 2 | 16 | 8 | 0.4 | |
| Luminal B | 30 | 3 | 5 | 10 | 12 | 0.6 | |
Column 1 represents the breast cancer subtypes by IHC and the rest of the columns represent the subtypes identified by centroid and random forest methods. The Error.rate column represents the difference in subtype classification according to IHC, and centroid and random forest methods, respectively.