Free Trial

Safari Books Online is a digital library providing on-demand subscription access to thousands of learning resources.

  • Create BookmarkCreate Bookmark
  • Create Note or TagCreate Note or Tag
  • PrintPrint
Share this Page URL
Help

6. Document Filtering > Alternative Methods

Alternative Methods

Both of the classifiers built in this chapter are examples of supervised learning methods, methods that are trained with correct results and gradually get better at making predictions. The artificial neural network described in Chapter 4 for weighting search results for ranking purposes was another example of supervised learning. That neural network can be adapted to work on the same problems in this chapter by using the features as inputs and having outputs representing each of the possible classifications. Likewise, support vector machines, which are described in Chapter 9, can be applied to the problems in this chapter.

The reason Bayesian classifiers are often used for document classification is that they require far less computing power than other methods do. An email message might have hundreds or even thousands of words in it, and simply updating the counts takes vastly less memory and processor cycles than training a neural network of that size does; as shown, it can be done entirely within a database. Depending on the speed required for training and querying, and on the environment in which it is run, a neural network may be a viable alternative. The complexity of a neural network also brings with it a lack of interpretability; in this chapter you were able to look at the word probabilities and s....


  

You are currently reading a PREVIEW of this book.

                                                                                                                    

Get instant access to over $1 million worth of books and videos.

  

Start a Free 10-Day Trial


  
  • Safari Books Online
  • Create BookmarkCreate Bookmark
  • Create Note or TagCreate Note or Tag
  • PrintPrint