Inferring Probability Models from Data

Abstract

One very useful way to draw conclusions from a dataset is to fit a probability model to that dataset. Once this is done, we can apply any of our procedures to the probability model to (for example) predict new data or estimate properties of future data. For example, you could flip a coin ten times and see five heads and five tails. To be able to estimate the probability of seeing heads on a future flip, you would need to fit a model. But once you had done so, you could also estimate the probability that five future flips would give you three heads and two tails, and so on.