In this chapter, we have learned how to cluster tweets using a variety of clustering methods. Though frequently touted as one of the most robust algorithms, we've shown that DBSCAN has problems with clustering tweets due to the nature of tweets being noisy. Instead, we see that older, more traditional methods, as well as a new method of clustering, would yield better results.
This points to a lesson—there is no one machine-learning algorithm to rule them all; there is no ultimate algorithm. Instead, we need to try more than one thing. In the chapters that follow, this theme will be more apparent, and we shall approach these with more rigor. In the next chapter, we will learn about basics of neural networks and apply them on handwriting to recognize digits.