Decision Trees are versatile Machine Learning algorithms that can perform both classification and regression tasks. In this post, we will learn how they use different cost functions to measure their performance depending on the type of problem they are solving.

Regression is a statistical technique that helps you to estimate the relationship between a dependent variable and one or more independent variables. It is basically the simplest machine learning algorithm you can imagine. For example, you can use regression to find out how the price of a house depends on its size, location, number of rooms, etc.

Backpropagation is a popular algorithm used in training neural networks, which allows the network to learn from the input data and improve its performance over time. It is essentially a way to update the weights and biases of the network by propagating errors backwards from the output layer to the input layer.

Have you ever had to analyze data with a complex structure, such as nested or hierarchical data, and found yourself struggling to find a suitable statistical model? If so, you might want to consider using linear mixed effects models.

I’ve been playing around with the Stan probabilistic programming language for full Bayesian statistical inference lately. Stan has nice R interface in the rstan package. This will be (hopefully) the start of a small serious where we fit some simple models using Stan.

Xgboost is short for eXtreme Gradient Boosting and currently one of best machine learning packages available for R. It is fast, memory efficient and can handle huge amounts of data and won several kaggle competitions.. Xgboost is now on CRAN and can be installed from there.