Machine Learning and Big Data with kdb+/q

CHAPTER 26Unsupervised Machine Learning: The Apriori Algorithm

Chapters on Linear Regression, Decision Trees, Random Forests, AdaBoost, Neural Networks and -Nearest Neighbours belong to the supervised machine learning algorithms family. In such cases, we are equipped with two types of variables in the data set: the -dimensional vector of features, , and, the -dimensional vector of dependent variables (usually ). Given the data set of observations of both features and corresponding dependent variables, we calibrate the chosen model using both types of data. The relationship between the features and the dependent variable is then learnt.

Our objective for this chapter is to assume that we have only features in our data set. The objective is to understand the relationship between features. This can be formalised as follows. The set of features , is generated by a joint density distribution ...