Blueprints for Text Analysis using Python: Machine Learning Based Solutions for Common Real World (NLP) Applications

What distinguishes this book from others in the same area is that all recipes included lead to an actual insight at the end, authors don't forget the so what behind the code and the analysis. Recipes are combined with some theory background too to help build intuition about the key concepts, which is useful. I come from an R background, but found the Python code easy to follow (use of many familiar data frame concepts).

I feel compelled to write this review to acknowledge the authors for their brilliant work. I am particularly impressed that this book is so well supported on github with colab links. The documentation and code is all so good.

I can't recommend this book highly enough. It provides an excellent foundation for understanding the fundamentals of NLP. I'd suggest starting here before moving onto the more complex Deep Learning techniques that are making lots of headlines these days. For many business applications you don't actually need state-of-the-art techniques - you need the fundamentals, and this book quite literally gives you the blueprints for applying them. In fact it also covers some more classical ML techniques and really useful concepts like explaining your classifier and word embeddings, so it's not just the basics. I particularly like the way that you can run the accompanying notebooks on Google Colab, meaning that you don't need to faff around setting up your environment and dealing with Python dependencies etc. All books should provide this option! Bravo to the author.

I noticed this in my bookshelf last week, and apparently I had bought it and forgotten to even take a look. I started skimming and was surprised at the usefulness of this book. It is very good. Pros: It does an excellent job of explaining fundamentals and common workflows of NLP. I know this, because my work is primarily NLP. It covers fundamentals quite well, such as tokenization, word vextors, similarity, classification; topic modeling, etc. It also gets into topic modeling. Later, there is a whole chapter about explainability of NLP models, which I am excited to read. I adore NLP insights. Cons: Like cookbooks, most things are blueprints. That’s very useful if you like those kinds of boxy explanations. I personally prefer typical book format, but this works too. I just find it a little distracting. But the book is rare in that it really explains the fundamentals. Many books junp straight to ML, or are only ML. This is good for foundation. It is also really useful and practical. This is now in my top four favorite NLP books. The pros absolutely outweigh the cons. And the datasets seem wonderful. I’m still reading and learning from this. Really glad I noticed I forgot to read.

Every chapter in this book is practical and can be replicated into your programming environment on your local machine at least. This book teaches us the powerful methods to analyse and process text data which will get at least 80% of the job done. Anyone with basic programming experience can benefit from this book. The best book to date on NLP.

I really enjoyed reading this book as it comprehensively and clearly explains the fundamentals of NLP, while also putting it directly into practice.

Connaître les méthodes de nlp