In this chapter, we’ve taken a look at some unique algorithms designed to remove unimportant features from a decision matrix. Although the scope of this chapter has been related directly to language classification, these types of algorithms can be incorporated into several unrelated fields. Various types of analysis in economics immediately come to mind, using the BNR algorithm to identify the more important data points of interest on noisy revenue or stock charts and perhaps even as a means of steganography detection. An implementation in general data mining could also apply to network event filtering or log mining.
In the next chapter, we’ll take a look at collaborative algorithms, which can be used to network different groups of users together to make collaborative filtering decisions.