The Art of Machine Learning

Learn to expertly apply a range of machine learning methods to real data with this practical guide.

Packed with real datasets and practical examples, The Art of Machine Learning will help you develop an intuitive understanding of how and why ML methods work, without the need for advanced math.

As you work through the book, you ll learn how to implement a range of powerful ML techniques, starting with the k-Nearest Neighbors (k-NN) method and random forests, and moving on to gradient boosting, support vector machines (SVMs), neural networks, and more.

With the aid of real datasets, you ll delve into regression models through the use of a bike-sharing dataset, explore decision trees by leveraging New York City taxi data, and dissect parametric methods with baseball player stats. You ll also find expert tips for avoiding common problems, like handling dirty or unbalanced data, and how to troubleshoot pitfalls.

You ll also explore:

• How to deal with large datasets and techniques for dimension reduction
• Details on how the Bias-Variance Trade-off plays out in specific ML methods
• Models based on linear relationships, including ridge and LASSO regression

• Real-world image and text classification and how to handle time series data
  Machine learning is an art that requires careful tuning and tweaking. With The Art of Machine Learning as your guide, you ll master the underlying principles of ML that will empower you to effectively use these models, rather than simply provide a few stock actions with limited practical use.

  Requirements: A basic understanding of graphs and charts and familiarity with the R programming language

  Autorentext

  Norman Matloff is an award-winning professor at the University of California, Davis. Matloff has a PhD in mathematics from UCLA and is the author of The Art of Debugging with GDB, DDD, and Eclipse and The Art of R Programming (both from No Starch Press).

  Klappentext

  "Teaches a range of machine learning methods, from simple to complex. Includes dozens of illustrative examples using the R programming language and real datasets. Covers not only how to use machine learning methods but also why these methods work and advice on how to avoid common pitfalls"--

  Inhalt
  Acknowledgments
  Introduction
  PART I: PROLOGUE, AND NEIGHBORHOOD-BASED METHODS
  Chapter 1: Regression Models
  Chapter 2: Classification Models
  Chapter 3: Bias, Variance, Overfitting, and Cross-Validation
  Chapter 4: Dealing with Large Numbers of Features
  PART II: TREE-BASED METHODS
  Chapter 5: A Step Beyond k-NN: Decision Trees
  Chapter 6: Tweaking the Trees
  Chapter 7: Finding a Good Set of Hyperparameters
  PART III: METHODS BASED ON LINEAR RELATIONSHIPS
  Chapter 8: Parametric Methods
  Chapter 9: Cutting Things Down to Size: Regularization
  PART IV: METHODS BASED ON SEPARATING LINES AND PLANES
  Chapter 10: A Boundary Approach: Support Vector Machines
  Chapter 11: Linear Models on Steroids: Neural Networks
  PART V: APPLICATIONS
  Chapter 12: Image Classification 
  Chapter 13: Handling Time Series and Text Data 
  Appendix A: List of Acronyms and Symbols 
  Appendix B: Statistics and ML Terminology Correspondence
  Appendix C: Matrices, Data Frames, and Factor Conversions
  Appendix D: Pitfall: Beware of “p-Hacking”!