Expected structure (some sections may come later, when the Qs are better understood):

• Questions: (sub-)questions themselves that we want to answer doing the research. The first one in there is a broad overall question, the following ones usually are more specific.
• Hypotheses: what hypotheses we are testing with the experiments
• Assumptions: about the system/question — e.g. what’s already available in terms of data, infra, hardware, etc.
• Possible approaches: may be filled after preliminary research
• Outcomes: what do we get out of answering the Qs

Example

Here’s a simplified example of a research question write up from a different project.

RQ1: Anomaly Detection

Questions

1. Can we identify periods in the systems’ monitoring data that represent anomalous behaviour? I.e. Ineffective usage
2. How much labelled data is required to effectively identify these anomalous periods?
3. Can the anomalous periods be identified across various configurations?
4. Can the same model be used for different system, or will this be a model per system?
5. Can we categorise the anomalies?

Some Qs are more specific then the others, they may cover different outcomes

Hypotheses

1. Given the time series data for a system we can correctly identify anomalous periods
2. Anomalous periods can be consistently identified across different system types
3. Provided labelled data, we can categorise identified anomalous periods into the following categories: A, B, C