Which step usually comes first in an ML pipeline?
Why D is the answer
Collecting and preparing data
Models depend on their data, so collection and preparation come first.
In detail
- A pipeline starts with collecting and preparing data.
- Most ML effort goes into data.
Use case
An illustrative scenario, not a real incident.
A team wants a demand model. It starts by collecting sales history and fixing gaps, because no model can recover from bad data.
See it in different ways
flowchart LR A["Collect data"] --> B["Prepare"] B --> C["Train"] C --> D["Evaluate"] D --> E["Deploy"] E --> F["Monitor"] F --> A
stateDiagram-v2 [*] --> Data Data --> Train Train --> Evaluate Evaluate --> Deploy Deploy --> Monitor Monitor --> Data: drift
flowchart LR S0["Collect and prepare data"] --> S1["Train"] --> S2["Evaluate"] --> S3["Deploy"] --> S4["Monitor"]
Key point: Pipelines start with data.
Why the other options are wrong
- A. Deploying the model
- Deployment comes later.
- B. Monitoring the deployed model for drift in production
- Monitoring comes after deployment.
- C. Retraining the model
- Retraining comes later.