Assuming I have time evolution of air temperature, air pressure, wind speed, wind direction, relative humidity and cloudiness, how can I predict weather based on this information?

I know that high air pressure increases chance of more concentrated sunlight and low air pressure increases chance of precipitation. I know that a decrease in temperature increases chance of snow or ice and also of low relative humidity and vice versa. I also know that higher wind speed means higher chance of a storm and that certain directions of wind can cause a change in temperature.

However I don't know what meteorological parameters I could predict with a combination of these variables. How would I predict the weather based on the time evolution of these variables?


1 Answer 1


You need a model. You could try to use Bayes theorem to build a model based upon conditional probabilities. See this reference from NOAA about weather forecasting: Probability Forecasting - Primer .

If I were to try to build a Bayesian model for weather forecasting, with the intention of using a limited data set of observations I had collected myself, I would look into existing software frameworks to see if I could use one of these. For example, the Infer.NET framework is available for free, and has several examples which would be a good starting point for learning about probabilistic modeling.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.