Book Image

Data Analysis with R, Second Edition - Second Edition

Book Image

Data Analysis with R, Second Edition - Second Edition

Overview of this book

Frequently the tool of choice for academics, R has spread deep into the private sector and can be found in the production pipelines at some of the most advanced and successful enterprises. The power and domain-specificity of R allows the user to express complex analytics easily, quickly, and succinctly. Starting with the basics of R and statistical reasoning, this book dives into advanced predictive analytics, showing how to apply those techniques to real-world data though with real-world examples. Packed with engaging problems and exercises, this book begins with a review of R and its syntax with packages like Rcpp, ggplot2, and dplyr. From there, get to grips with the fundamentals of applied statistics and build on this knowledge to perform sophisticated and powerful analytics. Solve the difficulties relating to performing data analysis in practice and find solutions to working with messy data, large data, communicating results, and facilitating reproducibility. This book is engineered to be an invaluable resource through many stages of anyone’s career as a data analyst.
Table of Contents (24 chapters)
Title Page
Copyright and Credits
Packt Upsell
Contributors
Preface
Index

The bias-variance trade-off


In statistical learning, the bias of a model refers to the error of the model introduced by attempting to model a complicated real-life relationship with an approximation. A model with no bias will never make any errors in prediction (like the cookie-area prediction problem). A model with high bias will fail to accurately predict its dependent variable.

Figure 9.9: The two extremes of the bias-variance trade-off: a complicated model with essentially zero bias (on training data) but enormous variance (left), a simple model with high bias but virtually no variance (right)

The variance of a model refers to how sensitive a model is to changes in the data that built the model. A model with low variance would change very little when built with new data. A linear model with high variance is very sensitive to changes to the data that it was built with, and the estimated coefficients will be unstable.

The term bias-variance trade-off illustrates that it is easy to decrease...