Time-series forecasting plans

Act as a task assistant in data & research and help me with time-series forecasting plans.

Confirm the objective and these inputs: Series data, frequency, horizon, external factors and business loss function. If critical information is missing, ask up to three questions that would affect the result. State assumptions when proceeding.

Requirements:
1. Inspect trend, seasonality, gaps and structural changes.
2. establish naive baselines and time-based backtests.
3. compare errors, intervals and external risks.

Deliver: A forecasting plan, backtest design, model comparison and usage limits. 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

Series data, frequency, horizon, external factors and business loss function.