I’m trying to predict the probability that a soccer shot in a video game will result in a goal based on input parameters regarding the shot’s trajectory (the player plays as the goalkeeper trying to block the shot).
The dependent variable is the percentage of the time a shot with certain input parameters has resulted in a goal.
I used linear regression to get an initial value, but encounter the issue that predictions lie outside [0,1] which isn’t possible for probabilities, obviously. I thought I could use this initial value as input to a logistic function, which would output values between 0 and 1, but this logistic function would have parameters that are also optimised using the goal percentages as the dependent variable, the same as the initial linear regression.
Is this an issue?
Beginner is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.