Hypothesis testing is the bedrock of inferential statistics, allowing us to draw inferences reliably about the population as a whole. Use this course to learn more about the distinction between descriptive and inferential statistics and how the latter seek to generalize from the sample to the population as a whole. Examine the components of a typical hypothesis test, such as the null and alternative hypothesis, the test statistic, and the p-value. You'll also explore type-I and type-II errors and the use cases and conceptual underpinnings of t-tests and ANOVA. By the time you finish this course, you will be able to identify use-cases for hypothesis testing and conceptually construct the appropriate null and alternative hypotheses for such tests.

In situations where two independent samples are drawn from different populations or where paired samples are available, such as in a before-after scenario, two-sample and paired T-tests are needed, respectively. Use this course to explore how two-sample T-tests can be used to test the null hypothesis that two independent samples have drawn from populations with equal means. You'll examine type I and type II errors and the use of paired samples T-tests. By the time you finish this course, you will be able to test whether two samples - either drawn independently or explicitly linked - are drawn from populations with equal means.

Two-sample T-tests are great for comparing population means given two samples. However, if the number of samples increases beyond two, we need a much more versatile and powerful technique - analysis of variance (ANOVA). Use this course to learn more about non-parametric tests and the ANOVA analysis. In this course, you'll explore the different use cases for Mann-Whitney U-tests, the use of the non-parametric paired Wilcoxon signed-rank test, and perform pairwise T-tests and ANOVA. You'll also get a chance to try your hand at the non-parametric variant of ANOVA - Kruskal Wallis test and post hoc tests, such as Tukey's honestly significant difference test (HSD). After completing this course, you will be able to account for the effect of one or two independent categorical variables, each having an arbitrary number of levels, on a dependent variable using ANOVA.

Understanding the bias-variance trade-off allows data scientists to build generalizable models that perform well on test data. Machine learning models are considered a good fit if they can extract general patterns or dominant trends in the training data and use these to make predictions on unseen instances. Use this course to discover what it means for your model to be a good fit for the training data. Identify underfit and overfit models and what the bias-variance trade-off represents in machine learning. Mitigate overfitting on training data using regularized regression models, train and evaluate models built using ridge regression, lasso regression, and ElasticNet regression, and implement ensemble learning using the random forest model. When you're done with this course, you'll have the skills and knowledge to train models that learn general patterns using regularized models and ensemble learning.

Classification models are used to classify or categorize data points into two or more categories. Learn how these models work and how you can evaluate your classification models using the confusion matrix and metrics such as accuracy, precision, and recall. During this course, you'll perform classification using both logistic regression and an imbalanced dataset. You'll also examine why precision or recall scores may be better metrics than accuracy to evaluate such models. Furthermore, build a classification model using decision trees, visualize the tree structure, and explore the variable importance assigned by this tree structure to understand and interpret the model. When you've finished this course, you'll be able to confidently use logistic regression and decision trees to build classification models and evaluate your models using accuracy, precision, and recall.

Clustering is an unsupervised learning algorithm that self-discovers patterns in data and helps identify logical groupings. Use this course to distinguish between supervised and unsupervised learning and recognize how regression and classification algorithms differ from clustering. Examine the basic principles of clustering models and how k-means clustering finds logical groupings in your data. Learn the evaluation techniques used in clustering and find the optimal number of clusters in your data using both the elbow method and the Silhouette score. Perform clustering on a dataset with multiple attributes and visualize clusters in your data using principal components. When you've completed this course, you'll be able to find groupings in your data using k-means clustering and compute the optimal number of clusters for your data.

Regression models are used to predict continuous values and are some of the most commonly used machine learning models. Use this course to grasp what exactly machine learning (ML) algorithms are and how you can use ML models to predict outcomes based on input data. Learn how regression models work, train them, and evaluate regression results using metrics such as R2 and RMSE. Perform regression analysis in R using the ordinary least squares regression. Build models using simple and multiple regression and train a regression model using cross-validation. Upon completing this course, you'll be able to perform regression to predict continuous values and evaluate these models using metrics such as the R2 and adjusted R2.

Hypothesis testing determines whether the educated guesses you've made about your data should be accepted or rejected. T-tests and ANOVA tests are some of the most commonly used methods in hypothesis testing. Knowing how to perform and interpret these tests are core data scientist skills. In this course, get hands-on running statistical tests on your sample data. Test assumptions made by statistical tests, run T-tests, perform ANOVA analysis, and interpret the results. Perform the one-sample t-test and the one-sample Z-test. Run the two-sample t-test to compare data from different groups or categories and the paired samples t-test to compare data across time. When you're finished with this course, you'll have the know-how to run and interpret statistical tests using the R programming language.

Statistical analysis involves making educated guesses known as hypotheses and testing them to see if they hold up. Use this course to learn how to apply hypothesis testing to your data. Examine the use of descriptive statistics to summarize data and inferential statistics to draw conclusions. Learn how population parameters differ from summary statistics and how confidence intervals are used. Discover how to perform hypothesis testing on sample data, construct null and alternative hypotheses, and interpret the results of your statistical tests. Investigate the significance of the p-value of a statistical test and how it can be interpreted using the significance threshold or alpha level. Additionally, examine the most commonly used statistical tests, the T-test and the analysis of variance (ANOVA). When you're done, you'll have the confidence to set up the null and alternative hypotheses for your tests and interpret the results.

Interpreting data is a core pre-processing step in data analysis and modeling. Use this course to practice using various dynamic statistical tools to explore and understand your data. During this course, you'll explore population distributions to model random variables, work with discrete and continuous probability distributions, and use discrete probability distribution types, such as the uniform, binomial, and Poisson distributions. You'll also examine continuous distributions, such as the normal and the exponential distributions. You'll round the course off by learning how to read and interpret QQ plots, which can be used to compare the distributions of two samples of data. When you're finished, you'll be able to use probability distributions to model events and understand your data.