Summary

In this chapter, we discussed the idea of the dimensionality of data. We went over why it could be useful to reduce the dimensionality of data and highlighted that the process of dimension reduction can reveal important truths about the underlying structure of data. We covered two important dimension reduction methods. The first method we discussed was market basket analysis. This method is useful for generating associative rules from complex data and can be used for the use case it was named after (analyzing baskets of groceries) or a wide variety of other applications (such as analyzing the clustering of survey responses). We also discussed PCA, a common way to describe data in terms of linear combinations of its dimensions. PCA is easy to perform with some linear algebra tools, and provides an easy way to approximate even very complex data.

In the next chapter, we will have a look at the different data comparison methods.