Frame a Problem for Machine Learning
Decide whether ML is needed and what to predict.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
Help me decide whether a machine learning model is the right tool. Business problem: [BUSINESS_PROBLEM]. Decision the prediction would change: [DECISION]. Data available: [DATA_AVAILABLE]. Current process and its accuracy: [CURRENT_PROCESS]. Work through these checks and mark each yes, no, or unknown: 1. Is there a clear target to predict, available at decision time, and recorded historically? 2. Is there enough labelled history, and does it represent the future? 3. Would a simple rule, a lookup table, or a better report do as well? 4. What is the cost of a false positive versus a false negative, in money or harm? 5. How will predictions be used and by whom, and can they act on them? 6. What accuracy would beat the current process, and how will we measure it? 7. Are there legal or fairness concerns (people affected by the decision) [VERIFY in [JURISDICTION]]? Output: a verdict (use ML, try a simple baseline first, or do not use ML) with reasons; the problem statement as "Given X at time T, predict Y for decision Z"; the baseline to beat; the evaluation metric and why; and a three-step plan with a go or stop decision after the baseline.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.