I don't understand why knowing the probability that an event occurs would change the model?
Hello,
This may be an easy question, but I would like to know how to incorporate the probability of an event in a linear (or non linear) equation.
For example, say I have the following linear equation: Y = (100X * W) + 20T + 10M - 50.
I collect data from an experiment, and find values X, W, T and M - no problem, I can now find Y. However, I found that W only has a 25% chance of occurring for any given experiment.
Can I (or how do I) incorporate this probability into my equation to produce a more accurate model?
Thanks!
I don't understand why knowing the probability that an event occurs would change the model?
I don't have emotions and sometimes that makes me very sad.
It would not change the model. It would change the distribution of the response. I would use Monte Carlo simulation. Just specify the distribution or probabilities of your different inputs, run the simulation and evaluate the distribution of the ouput. Follow up with a sensitivity analysis to determine which inputs drive the most variation in output.
mjgarrin (12-12-2016)
Thanks for the tips. I wasn't sure if I was able to do that or not. I was using the following model Y = (100X * [W*0.25]) + 20T + 10M - 50 to represent the 25% probability of W, but you were right, this changes my response. I removed this, and performed the monte carlo simulation and found the error in my model to be about 0.4%.
You mentioned performing sensitivity analysis...how do I do this? I cannot seem to find any simple tutorial online to perform this?
Thank you.
Tweet |