Naïve Bayes classifiers

The Naïve Bayes classifier has a strict requirement: the features must be independent (that is, conditional dependence between features is null). It also restricts its applicability. The Naïve Bayes classification is better understood through simple, concrete examples [5:5].

Introducing the multinomial Naïve Bayes

We illustrate the Naïve Bayes classification in the context of predicting the fluctuation of the interest rate of treasury bills.

The first step is to list the factors that potentially may trigger or cause an increase or decrease in the interest rates. For the sake of illustrating Naïve Bayes, we select the consumer price index (CPI), change in the federal fund rate (FDR), and the growth domestic product (GDP) as a first set of features. The terminology is described in the Terminology section under Finances 101 in the Appendix.

The use case is to predict the direction of the change in the yield of the 1-year Treasury bill (1yTB), considering the change in the...