Practical Probabilistic Programming cover
welcome to this free extract from
an online version of the Manning book.
to read more

About this Book


Lots of decisions, whether in business, science, the military, or everyday life, involve judgment calls under uncertainty. When different factors sway you in different directions, how do you know what to pay attention to the most? Probabilistic models enable you to express all the relevant information about your situation. Probabilistic reasoning lets you use these models to determine the probability of the variables that make the most difference to your decision. You can use probabilistic reasoning to predict the things that are most likely to happen: will your product be a success at your target price; will the patient respond well to a particular treatment; will your candidate win the election if she takes a certain position? You can also use probabilistic reasoning to infer the likely reasons behind what happened: if the product failed, is it because the price was too high?

Probabilistic reasoning is also one of the main approaches to machine learning. You encode your initial beliefs about your domain in a probabilistic model, such as the general behavior of users in response to products in your market. Then, given training data, perhaps about the response of users to specific products, you update your beliefs to get a new model. Now you can use your new model to predict future outcomes, like the success of a planned product, or infer likely causes of observed outcomes, like the reasons behind the failure of a new product.


About the code and exercises