Customer Reviews

Semisupervised Learning for Computational Linguistics (Chapman & Hall/CRC Computer Science & Data Analysis)

5 star:		(1)
4 star:		(1)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

We're finally getting to the point where Computational Linguistics will start to see their titles in the titles. In the past one would have to piggyback off of another discipline to get the information they needed. This book is a must for anyone learning anything statistical in the NLP field. I took a class which covered nearly all of the topics in this book just months before the book came out. I struggled through some of the concepts and spent many a sleepless night going over an academic paper at least one more time getting those concepts down. On the last day of class the professor suggested this new title. I went and bought and most of the hard stuff I had struggled with solidified in my mind. A great feeling! I wish it was the textbook.
About the book itself; it does assume the reader is pretty math savvy. Some sections claim they are not breaking down a proof even though the only thing on the page are equations. But on the flip side, Abney does a fantastic job of grounding the terminology before launching into that. The first few chapters are very informative and patient with the reader. It is also excellent if you just need a refresher on any of these topics.

The book is about introduction to semi-supervised methods from machine learning applied on text processing. Author did a great job introducing the topics of machine learning and goes chapter by chapter to the goal - the need and motivation for semi-supervised learning.

A nice thing is that much of mathematical rigor is removed but instead replaced by (quite lengthy) explanations. Many machine learning concepts are introduced on the fly which is great for students new to the field.
Unsupervised and Supervised methods are presented gently and in a clear way such that for somebody with not so strong linear algebra will be still able to grasp.

On the downside, his english is bit hard to read. I am used to reading scientific publications and general english texts, but his language sometimes forced me to re-read his sentences two or three times to understand his explanations.