Lightweight Machine Learning Classics with R
DRAFT v0.2 2020-05-11 22:28 (16b9c8d)
This is a draft version (distributed in the hope that it will be useful) of the book Lightweight Machine Learning Classics with R by Marek Gagolewski.
Copyright (C) 2020, Marek Gagolewski. This material is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND 4.0).
You can access this book at:
- https://lmlcr.gagolewski.com/ (a browser-friendly version)
- https://lmlcr.gagolewski.com/lmlcr.pdf (PDF)
- https://github.com/gagolews/lmlcr (source code)
Aims and Scope
Machine learning has numerous exciting real-world applications, including stock market prediction, speech recognition, computer-aided medical diagnosis, content and product recommendation, anomaly detection in security camera footages, game playing, autonomous vehicle operation and many others.
In this book we will take an unpretentious glance at the most fundamental algorithms that have stood the test of time and which form the basis for state-of-the-art solutions of modern AI, which is principally (big) data-driven. We will learn how to use the R language (R Development Core Team 2020) for implementing various stages of data processing and modelling activities. For a more in-depth treatment of R, refer to this book’s Appendices and, for instance, (Wickham & Grolemund 2017, Peng 2019, Venables et al. 2020).
These pages contain solid underpinnings for further studies related to statistical learning, machine learning data science, data analytics and artificial intelligence, including (Bishop 2006, Hastie et al. 2017, James et al. 2017). We will also appreciate the vital role of mathematics as a universal language for formalising data-intense problems and communicating their solutions. The book is aimed at readers who are yet to be fluent with university-level linear algebra, calculus and probability theory, such as 1st year undergrads or those who have forgotten all the maths they have learned and need a gentle, non-invasive, yet rigorous introduction to the topic. For a nice, machine learning-focused introduction to mathematics alone, see, e.g., (Deisenroth et al. 2020).
I’m currently a Senior Lecturer (equivalent to Associate Professor in the US) in Applied AI at Deakin University in Melbourne, VIC, Australia and an Associate Professor in Data Science at Warsaw University of Technology, Poland. My main passion is in research – my primary interests include machine learning and optimisation algorithms, data aggregation and clustering, statistical modelling and scientific computing. Explaining of things matters to me more than merely tuning the knobs so as to increase a chosen performance metric (with uncontrollable consequences to other ones); the latter belongs to technology and wizardry, not science.
I’m an author of more than 70 publications. I’ve developed several open source R and Python packages, including stringi, which is among the most often downloaded R extensions.
On top of that, I teach various courses related to R and Python programming, algorithms, data science and machine learning – and I’m good at it. This book was also influenced by my teaching experience at Data Science Retreat in Berlin, Germany.
This book has been prepared with pandoc, Markdown and GitBook. R code chunks have been processed with knitr. A little help of bookdown, good ol’ Makefiles and shell scripts did the trick.
The following R packages are used or referred to in the text: bookdown, Cairo, fastcluster, FNN, genie, gsl, ISLR, keras, knitr, Matrix, microbenchmark, pdist, RColorBrewer, recommenderlab, rpart, rpart.plot, rworldmap, scatterplot3d, stringi, tensorflow, tidyr, titanic, vioplot.
During the writing of this book, I’ve been listening to the music featuring John Coltrane, Krzysztof Komeda, Henry Threadgill, Albert Ayler, Paco de Lucia and Tomatito.