Python Machine Learning By Example - Second Edition

Random forest (or decision tree)

Linear separability of data does not matter to the algorithm. And it works directly with categorical features without encoding, which provides great ease of use. Also, the trained model is very easy to interpret and explain to non-machine learning practitioners, which cannot be achieved with most other algorithms. Additionally, random forest boosts decision tree, which might lead to overfitting by ensembling a collection of separate trees. Its performance is comparable to SVM, while fine-tuning a random forest model is less difficult compared to SVM and neural networks.