Hello, everyone. This is. I'm a teaching assistant for the Classroom Management course. Today, I'm glad to give you a brief introduction to ClusterEnG, which is a tool that allows you to play with different classroom algorithms you have learned from the course. First of all, what is ClusterEnG? ClusterEnG are to developed by the Computational Biology Group at UIUC. It allows the users to perform ClusterEnG analysis in an interactive way. On ClusterEnG, you can create your own data, choose different ClusterEnG algorithms, and the results. We recommend you to play with it, because it can help you understand how different the ClusterEnG algorithms work in practice. To use ClusterEnG you will first need to visit a website which is provided here. After opening ClusterEnG, you will see that it includes three major steps. The first step is data preparation. And second step is choosing the Algorithms, and the parameters that you want to use. And final step is the visualization of the cluster results. In the rest of this tutorial, we'll introduce the details of those three steps. In the data preparation step, you need to prepare the data that you want to perform classroom analysis on. There are two choices. The first one is to upload the own data file that you have. The uploaded file could either be in CSV format where the elements are separated by commas. Or the TXT format where the elements are separated by spaces which helps. To give you a concrete example, here I have created very simple CSV file that includes 15 data points, and each data point is two dimensional. After choosing the created data file, I just need to hit the start button to upload it. And after uploading, I choose rows here, because all the little points are stored as rows in the uploaded file. Also, I choose no here, because I don't have row names or IDs in my first data file. Another option for preparing the data is to use the sample data size provided by clustering. Once you're done with preparing the data, next step is choose the clustering algorithm and the parameters that you want to use. As you can see from the website, clustering actually covers a quite wide range of different clustering algorithms including k-means, k-medoids, hierarchical clustering. DBSCAN and several others. If you are not sure which clustering algorithm you should use, there's actually a tutorial page on the website that compares the different pros and cons of different clustering algorithms. You can read through them and choose the appropriate algorithm based on your needs. For our simulator, lets just use k means as our class algorithm and that sets the number of class rows to three. Then we hit submit to round k means on the uploaded file. The final step is to initialize and interpret the classroom results. Since classroom is designed to support both low dimensional and high dimensional data class range, we visualize the results using PCA which is a very well known dimension reduction technique. There are two different models, the first one is two-dimensional models Which provides the clustering plots on two of the first rate principal components. And the second mode is the three-dimensional mode which plots the visualization results on the first grid principal components. Since the data file we just used is two-dimensional, we only need to look at the first plot under the two-dimensional viewing mode. We can see the 15 data points are separated into three different clusters by the k-means algorithm, where the different clusters are represented using different colors We're now introducing the basic usage of ClusterEnG. If you want to know more about this tool you can find more resources on their website. Also in the following weeks we'll go into the detail of several different clustering algorithms by [INAUDIBLE]. Thank you for watching. [SOUND] [MUSIC] [SOUND]