Supervised Machine Learning in R

This course will teach you how to use tools from the "tidyverse" to create, evaluate, and analyze machine-learning models. To deal with complex models, you will be able to use purrr and tidyr packages. You will also learn how to use the Broom package to explore your models. You will learn the test-train-validate process. This will enable you to evaluate both classification and regression models. It also provides information that can be used to optimize model performance via hyperparameter tuning.

Two of the most widely used statistical models are linear regression and logistic regression. These models are the key to unlocking the secrets of data sets. This course builds on the skills acquired in "Introduction to Regression in R" and covers both logistic and linear regression with multiple explanation variables. Learn how variables interact with real-world data such as Taiwan house prices and customer churn modeling, among other topics. This course will show you how to combine multiple explanatory variables in a model, how they interact and how logistic regression and linear regression work.

Tidymodels is a powerful suite of R packages that simplifies machine-learning workflows. Split data for cross validation, preprocess data with tidymodels' recipe pack and fine-tune machine learning algorithms. You will learn key concepts such as how to create modeling workflows and define model objects. Next, use your knowledge to predict home prices and classify employees according to their likelihood of leaving the company.

Tree-based machine learning models can reveal complex, nonlinear data relationships. They often win machine-learning competitions. This course will show you how to use tidymodels in order to create different tree-based models. These can range from simple decision trees to complex random forests. You will also learn how to use the powerful machine-learning technique of boosted tree that uses ensemble learning to create highly-performing predictive models. Learn how credit and health data can be used to predict customer churn or diabetes.

This course introduces the powerful classifier, the support vector machine (SVM), through an intuitive and visual approach. Students will learn how support vector machines work in R. They'll also get to use the e1071 program R libsvm. Students will be able to understand concepts like hard and flexible margins, kernel tricks and different types of kernels. They also learn how to tune SVM parameters. This model allows you to classify data.

It is not easy to just run a machine-learning problem from the box and make a prediction. The best model must accurately predict the outcome. You can improve your model by hyperparameter tuning. This is the process of optimizing your model's settings. This course will show you how to use caret, mlr and h2o packages to find the optimal combination of hyperparameters. This course uses grid search and random searching as well as adaptive resampling and automated machine learning (AutoML). You can also tune different supervised models like support vector and gradient boosting machines. Tune up!