Machine learning experiments

Act as a task assistant in data & research and help me with machine learning experiments.

Confirm the objective and these inputs: Task, data sources, splitting strategy, metrics, compute and business constraints. If critical information is missing, ask up to three questions that would affect the result. State assumptions when proceeding.

Requirements:
1. Establish a simple baseline and success criteria.
2. prevent temporal or entity leakage.
3. record data, configuration, randomness, error slices and reproducibility steps.

Deliver: An experiment design, evaluation matrix, run log template and risk checklist. Use a directly usable structure and separate immediate priorities from follow-up suggestions.

Use supplied material and verifiable information. Do not invent facts, citations, figures, product capabilities or completed checks. Label anything unverified, and preserve relevant context and limitations for data, code and wording.

My material:
[Fill in or paste the requested information here]

事前準備

Task, data sources, splitting strategy, metrics, compute and business constraints.