Introduction

The first two recipes deal with summarizing a series of data. For example, assume someone asks, "How old is everyone?". A valid response could be to enumerate through the age of each person, but depending on the number of people, this could take minutes if not hours. Instead, we can answer in terms of the average or in terms of the median to summarize all the age values in one simple number.

The next two recipes are about approximating an equation that most closely fits a collection of points. Given two series of coordinates, we can use a linear or quadratic approximation to predict other points.

We can detect relationships between numerical data through covariance matrices and Pearson correlation calculations as demonstrated in the corresponding recipes.

The Numeric.Probability.Distribution library has many useful functions for deeper statistical understanding as demonstrated in the Bayesian network and playing cards recipes.

We will also use Markov chains and n-grams for further...