R Tutorial: Introduction to classification trees

Want to learn more? Take the full course at https://learn.datacamp.com/courses/machine-learning-with-tree-based-models-in-r at your own pace. More than a video, you'll learn hands-on coding & quickly apply skills to your daily work. --- Trees have advantages and disadvantages like any other algorithm. There are several advantages that make decision trees a useful machine learning algorithm. One of the biggest advantages of decision trees is that both the training and prediction process is easy to explain. They are even easier to explain than a linear model! Understanding why a particular prediction was made is as simple as following a path down a tree. Trees can be displayed visually, and are easily interpreted even by a non-expert (especially if they are small). Decision trees require no normalization or standardization for numeric features. Trees can also handle categorical features without the need to create dummy binary indicator variables (or to use other types of categorical encodings). Some tree software implementations make use of a technique called "group splitting" to handle categorical features within the partitioning algorithm itself. Decision trees also handle missing data in a elegant form, though not all decision tree software supports native handling of missing data. One popular method for dealing with missing values in trees is when you arrive at a split point and the feature value is not available, you can choose to go right or left at random, and then proceed as normal. Another method involves going down both branches at the split with missing data, and then when you finally reach a leaf node in each path, you average the leaves together for the final prediction. Since no numeric transformations or categorical encodings are required, missing values can often be handled automatically and since they are robust to outliers, trees require little to no data preparation. Decision trees recursively partition the feature space into su