- Building Smart Web 2.0 Applications
- Toby Segaran
- Web Applications
If you’ve ever wondered how web-sites can suggest other books, music or movies that you might enjoy then this book will tell you. It’s packed with code for grouping, clustering, filtering and analysing information. If you use Python and you either understand, or want to learn more, about Bayesian and Decision Tree Classifiers, Neural Networks, Support-Vector Machines, k-Nearest Neighbours, Clustering, Multidimensional Scaling, Non-Negative Matrix Factorisation and Optimisation then this book could well be for you.
But, if you’re like me and use Ruby and aren’t a maths wizard then most of the content may well pass at great height over your cranium. Yes, I did read it all, if only to try and glean more of an understanding of the techniques involved. It started well, discovering groupings in RSS feeds, then the concepts became trickier to grasp. I think the only thing that kept me going was one of the last chapters on evolving intelligence. Being a developer of software I’ve often wondered how you can write something, given specific rules, that can develop it’s own intelligence. Especially when these things can play games with themselves and learn as they go.
The only chapter that I’m certainly going to re-read and look into is the one on discovering groups. My idea is to add a related posts, and possibly even unrelated posts, section under each post on this web-site. From what the book says this would look at the text of each post and find ones with similar words and counts of words. I may even be able to plot a dendrogram showing clusters of topics.