Practical Linear Regression in R for Data Science in R

Explore why to use R and RStudio. Discover that R is an open source, cross-platform language for statistical computing and graphics, with GUI options, extensible packages, and reproducible research.

Install R and R studio on your PC by downloading R for Windows from the Quraan project page, run the installers, and consult manuals and frequently asked questions when needed.

Explore the RStudio interface—source, console, environment, history, and files panes—and learn to set up a working directory, access help, execute code, and view plots.

Import data into R from Excel, CSV, and text files by setting the working directory, then read data using read.table, read.csv, and read_excel with header and separator options.

Master the basics of R by installing and managing packages, exploring data types, vectors, factors, and data frames, and learning how to read data and write simple scripts.

Learn how to install and manage R packages in RStudio, including loading from repository or archive data, using install.packages and library, with script annotations.

Learn how to assign variables in R, perform arithmetic such as 8 plus 7 and 4 times 2, and print results in the console with meaningful names.

Explore the core data types and data structures in R, including numeric, character, integer, logical, and complex vectors, plus attributes, class, and factors with levels.

In this lab, learn to assign numeric and integer variables in R, convert types with as.integer and as.character, and create and order factor variables with levels.

Explore vectors and vector operations in R, including type conversion, indexing, combining vectors, elementwise arithmetic, recycling rules, and built-in functions like max and variance.

Explore data frames, the core two-dimensional tabular data type with name, size, weight, and four observations; use structure and summary to view types (name as factor; size and weight numeric).

Explore control structures in R to manage script flow, master if, if else, and while statements, and learn vectorized alternatives and basic output with print.

Explore regression analysis as supervised learning, focusing on linear regression, its formula and error term, and use scatter plots, box plots, and density plots to assess model fit in R.

Explore building a linear regression model in R with the diamonds dataset. Use scatterplots to relate price to carat, then check outliers and normality with box plots and density.

Assess whether your regression model in R best fits the data by examining metrics like R-squared, adjusted R-squared, F-statistics, standard error, p-values, and information criteria such as AIC.

Explore linear regression diagnostics by evaluating a simple regression of diamond price on carat, checking model significance, coefficient p-values, and R-squared and adjusted R-squared for fit.

Explore AIC and BIC, Akaike and Bayesian information criteria, as measures of model fit and tools for selecting among regression models, with lower values indicating better fit.

Evaluate a regression-based prediction model with root mean square error, compare in sample with out of sampling error to assess generalization, and note the confusion matrix for classification.

Predict diamond prices with a full linear regression model using all variables and the predict function, then assess rmse as in-sample error.

Split the data into training and test sets (80/20) with a seed for reproducibility, train a regression model, and evaluate it with out-of-sample RMSE to ensure robust performance.

Build a multiple linear regression model in R with YouTube, Facebook, and newspaper advertising to predict sales; split data, estimate coefficients, assess significance, and refine.

Predict sales with a multiple linear regression (additive) model using Facebook, YouTube, and newspaper campaigns, validate predictions with error metrics, and examine additive versus interaction effects across channels.

Learn how to handle categorical predictors in linear regression using dummy coding in R, including creating and interpreting dummy variables, contrast matrices, and the impact of coding on coefficients.

Explore how anova handles categorical variables with more than two levels in linear regression in R, using dummy coding and contrasts to interpret salary data.

Train a logistic regression model to classify rocks versus mines with sonar data, then evaluate accuracy using a confusion matrix at a 0.5 threshold, reporting sensitivity and specificity.

Analyze the receiver operating characteristic curve and AUC to balance true positive rate and false positive rate across thresholds, and interpret the area under the curve for model performance.

This bonus lecture invites learners to explore the instructor's Udemy page and Geo World channel for guided paths in JS, remote sensing, data science, and machine learning.