| Definition |
It learns by utilizing labeled data. |
It is trained on unlabeled data. |
The practitioner interacts with its environments, were as performing actions and leaning from the rewards or errors. |
| Problem type |
Classification and regression. |
Clustering and association. |
Based on rewards. |
| Data type |
Labeled data. |
Unlabeled data. |
Absent of predefined data. |
| Training |
External supervision. |
No supervision. |
No supervision. |
| Approach |
Inputs being labeled are mapped to the known output. |
Patterns are being understood and output is discovered. |
Trial and error are being implemented. |
| Operation |
ML. |
ML. |
ML. |
| Exploration |
No exploration. |
No exploration. |
Adapts through exploration. |
| Strategy |
Learning algorithm and data-dependent. |
Classification and data-dependent. |
Learns from experience. |
