As companies and researches rush to implement more and more machine learning practices into their organizations, occasionally they sacrifice understanding the complexities of statistics practices in order to achieve results faster. …

Did you come here expecting to find the “Hundred-page machine learning book” or “Elements of statistical learning”? This article is going to be a bit different. In this article, I want to highlight the 5 books that expose the controversial policies and business models, as well as the surveillance abuses…

Lasso is a regularization constraint introduced to the objective function of linear models in order to prevent overfitting of the predictive model to the data. The name Lasso stands for Least Absolute Shrinkage and Selection Operator. It turns out that the Lasso regularization has the ability to set some coefficients…

Recursive feature elimination (RFE) is the process of selecting features sequentially, in which features are removed one at a time, or a few at a time, iteration after iteration. Given a machine learning model, the goal of recursive feature elimination is to select features by recursively considering smaller and smaller…

Mutual information (MI) is a non-negative value that measures the mutual dependence between two random variables. The mutual information measures the amount of information we can know from one variable by observing the values of the second variable. The mutual information is a good alternative to Pearson’s correlation coefficient, because…

A common step before training machine learning algorithms is the discretization of continuous variables. But, why do we discretize continuous variables and how do we sort continuous data into discrete values? These are the questions that we will address throughout this article. Feature engineering for machine learning This article is the fifth in a series…

You’ve probably heard that before training machine learning models, data scientists transform random variables to change their distribution into something closer to the normal distribution. But, why do we do this? Which variables should we transform? Which transformations should we use? …

You’ve probably done your online searches on “Feature Selection”, and you’ve probably found tons of articles describing the three umbrella terms that group selection methodologies, i.e., “Filter Methods”, “Wrapper Methods” and “Embedded Methods”. Under the “Filter Methods”, we find statistical tests that select features based on their distributions. These methods…