Hi. I am Hong Lay Drong. I am teaching assistant of this Coursera course. Today, I'm going to introduce how to use an online clustering tool to apply an agglomerative hierarchical clustering algorithm on real data sets. The technical details of the hierarchical clustering algorithms already introduced in previous slides. If you're not familiar with the technical details, you may want to review the previous lessons. There are several different ways to implement the hierarchical clustering algorithm. The major difference is on how do you define the dissimilarity. And to clarify in this system, we use average link definition of dissimilarity. In this example, we're going to use the data set of gene expression. You can find the list of such data sets in the link given in this slide. In this data set, each row represents a gene, and each column represents a sample. The value in this data set represents the measured gene expression of a certain gene within the surgeon sample. To start with, we select the surgeon gene expression data set and insert the name of the data set. Here, we use an example of GDS10 data set. Click on submit. Here is how a hierarchical clustering result looks like. This part visualized the clustering results of genes represented by a dendrogram. If you want to get a clustering results with a certain clustered number, let's say, if you want to get a clustering results with four clusters, we can cut at a certain point of the dendrogram and get clustering results. Similarly, here's a dendrogram of samples. And a heat map visualize a clustering data set. And a larger value is represented by green color and smaller value is represented by red color. On the website, you can actually interact with the clustering results. As you can see, if I want to determine a certain branch, the dendrogram, we can select a certain branch, click and hold, To determine and see more details on the clustering results. I'm going to zoom out, you can simply click on this unzoom rows button, To get to the original results. If you want to perform hierarchical clustering on your own data set, you can also do so on the website. Upload the data set here. This is the data set of restaurant locations in Waterloo, Canada. Choose hierarchical clustering as a clustering method and as I said, the number of clusters here. Click on submit. The visualization of the clustering results is going to be shown here. Of course, you can also try different parameter settings or try these algorithms on different real data sets to see how the results differ. If you're interested in exploring more on hierarchical clustering algorithm, here are some open questions. Thank you, and enjoy. [MUSIC] [MUSIC] [SOUND]