The best way to calculate a risk score based on driving conditions

24 Views Asked by At

If I were to have a list of current weather conditions, including measurements like temperature, humidity, etc. What would the mathematical best way to calculate a risk score for bad driving conditions based on those measurements?

1

There are 1 best solutions below

0
On

Get a dataset with different observations of those weather conditions and a risk score for bad driving conditions, perhaps from a human assessor or from the weather service.

Then estimate a statistical model that predicts those risk scores based on weather conditions. It could be a Logit/Probit model if your risk assessment is just binary (bad driving conditions vs good driving conditions), or ordered Logit if you have more than two categories. OLS might also be used. OLS would also have a precise meaning to "mathematical best way" - the statistical model would fit the data best in a least squares sense.