Unsupervised learning and big data

The previous section illustrated how a deep neural network could be used to classify a limitless supply of input images as an instance of everyday creatures or objects. From another perspective, one might also understand this as a machine learning task that takes the highly dimensional input of image pixel data and reduces it to a lower-dimensional set of image labels. It is important to note, however, that the deep learning neural network is a supervised learning technique, which means that the machine can only learn what the humans tell it to learn—in other words, it can only learn from something that has been previously labeled.

The purpose of this section is to present useful applications of unsupervised learning techniques in the context of big data. These applications are in many ways similar to the techniques covered in Chapter 9, Finding Groups of Data – Clustering with k-means. However, where previous unsupervised learning...