During the research process, there may be a need to build multiple models and tracking how they differ in terms of inputs/outputs/assumptions/metrics/etc is vital for interpretation and traceability of model results. Models will always be conditional on the data used and the assumptions made, and these conditions must be made clear.
Each model should be documented in more detail via the Model Report template, but documentation of the Modelling Strategy should be completed first to clarify the high-level process and gain support from the technical reviewer. It is important to think carefully about the appropriate techniques and how they will be assessed for performance before beginning the modelling activity. This document aims to give advice on what needs to be considered in the modelling strategy. This has already been given some thought in the project scoping but following data acquisition and exploration may need updating.
Key points
The modelling strategy informs the modelling activities before they begin, and ties together the model reports produced in each iteration
Should be completed before any modelling activities takes place
Considers the higher-level concerns such as who needs to be involved, what tools are required, what the methodology is and how the outputs will be exploited
Answering every prompt is not necessary, but you should consider each one and if they are relevant to your project/proposal
Similarly, the list of prompts is not necessarily exhaustive, add more detail if and when you feel it is appropriate