Machine Learning Techniques for Text: Apply modern techniques with Python for text processing, dimensionality reduction, classification, and evaluation

by Nikos Tsourakis

Take your Python text processing skills to another level by learning about the latest natural language processing and machine learning techniques with this full color guide With the ever-increasing demand for machine learning and programming professionals, it's prime time to invest in the field. This book will help you in this endeavor, focusing specifically on text data and human language by steering a middle path among the various textbooks that present complicated theoretical concepts or focus disproportionately on Python code. A good metaphor this work builds upon is the relationship between an experienced craftsperson and their trainee. Based on the current problem, the former picks a tool from the toolbox, explains its utility, and puts it into action. This approach will help you to identify at least one practical use for each method or technique presented. The content unfolds in ten chapters, each discussing one specific case study. For this reason, the book is solution-oriented. It's accompanied by Python code in the form of Jupyter notebooks to help you obtain hands-on experience. A recurring pattern in the chapters of this book is helping you get some intuition on the data and then implement and contrast various solutions. By the end of this book, you'll be able to understand and apply various techniques with Python for text preprocessing, text representation, dimensionality reduction, machine learning, language modeling, visualization, and evaluation. This book is for professionals in the area of computer science, programming, data science, informatics, business analytics, statistics, language technology, and more who aim for a gentle career shift in machine learning for text. Students in relevant disciplines that seek a textbook in the field will benefit from the practical aspects of the content and how the theory is presented. Finally, professors teaching a similar course will be able to pick pertinent topics in terms of content and difficulty. Beginner-level knowledge of Python programming is needed to get started with this book.

448 pages, Paperback

Published October 31, 2022

Reviews

[Manny · Rating 4 out of 5]

I shared an office with Nikos for years, and we wrote a bunch of joint papers, where often I would contribute something based on classical computational linguistics methods and Nikos would contribute something based on machine learning; I think the one I'm most pleased with is this effort from 2017. Nikos knew the machine learning toolkits very well, and he was always able to get things done with great ease, so I never got around to acquiring these useful skills. Now I've just left the University of Geneva and moved to the University of South Australia, and I no longer have direct access to his expertise; but, with excellent timing, Nikos has published this book. It's almost like having him back on the other side of the office again.

Some books on machine learning are full of matrix algebra and partial derivatives, but there's little of that in Nikos's book: he's a hands-on kinda guy, and the book is constructed in a hands-on kinda way. He organises the text around ten case studies, where each time you want to do some kind of data analysis task which involves machine learning, and he walks you through a solution using scikit-learn, pytorch, keras, pandas, matplotlib and the other Python libraries. The examples are engaging: detecting spam emails, classifying newgroup posts, recommending music titles, and the like. By the end, you're using advanced deep learning techniques to do things including building nontrivial chatbots.

It's amazing how powerful and versatile these packages have become. You just type 'pip install' plus the name of the package, and two minutes later you have something sitting on your machine which was cutting edge software only ten years ago, in some cases less than that: you can organise data, train a model, apply it, visualise the results. The packages are neatly set up with easy-to-use interfaces that let you do everything with a few simple commands. Nikos walks you through it, and you see how straightforward it is once you've mastered the tricks. These days (the book came out just before ChatGPT), it's become even easier: once you know something is possible, you can ask Chat for the invocations, and it'll generally be able to give you a solution that either works, or is close and can be fixed after a bit of discussion. The possibilities it opens up are staggering.

Many times, I was reminded of Flynn's thought-provoking book What Is Intelligence? , which I read back in 2010. Flynn discovered the effect that bears his name, according to which IQ scores steadily rise by a few points a decade. People have questioned the validity of the Flynn Effect: are we really getting smarter at that rate, how can such a thing be possible? Flynn argues persuasively that it's actually not mysterious at all. A large part of intelligence, he says, is about collecting together a better mental toolkit. As time progresses, more and more useful tools are developed, and it's easier and easier to acquire them. Machine learning is a striking example. Not long ago, being able to use these techniques would have made you an exceptionally smart person. Now, it's just a bunch of tricks that anyone can pick up with a little diligence. But the tricks are no less powerful just because they're easily accessible. Get Nikos's book or a similar one, read through it, experiment a bit under ChatGPT-4's friendly guidance, and you'll feel measurably smarter. Maybe you'll actally be measurably smarter. Why don't you find out?

[Kostas Kampourakis · Rating 5 out of 5]

Clear and accessible for non-experts.