 Welcome back to learning analytics tools course. In this week, we will continue discussing about diagnostic analytics and algorithms in that especially clustering. So, what is clustering? Clustering is grouping a similar behavior or the items which has a similar behavior in a clusters. For example, if there are say 100 students taking class and the behavior based on the behavior, the background profile, you can group them. So, that is called clustering, grouping the students based on their similarity in behavior, especially this behavior is different from the other group. So, if you have two groups, the student in group one will have a different behavior from the group two, that is how you create the clusters. In learning environments, we can use students' interactions with the system, the numerical data to create clusters. So, you have to convert the data into frequency, the time span, those kind of informations, then that can be used to cluster. Also, we can use the action sequences to cluster, but that is not discussed in this course. So, let us see how to use the interaction behavior in the terms of numerical values can be clustered. Before that, let us discuss what is clustering and what are the famous clustering algorithms. There are two major category of clustering algorithms. There are more, but the two are very, very important major. The one is based on centroid, you might have seen the clusters, they are creating using centroids. For example, in this picture, you see there are three clusters. This picture is credited from Wikipedia. So, here is the link. In this week, we will see example algorithm for one centroid based and one example algorithm for the connectivity based. So, before we jump into what is clustering and the algorithms in clustering in the two main categories, can you think of why we need clustering for data from learning environments? Like you can choose any learning environment, like a classroom environment or a MOOC or TALE, anything. So, choose one of the learning environment and list down at least two applications of where clustering will be useful, why we have to do clustering. It is not that I want to do cluster based on students interaction with MOOC, but why, what is the information you want to use after creating clusters. Please pass this video after you will sit down and assume to continue. So, the clustering of learners based on the interaction. For example, students interacting with MOOC or TALE based on the interaction you can cluster, then you can use that to provide a feedback. So, instead of providing same feedback to all learners, you know this learners vary by three or four clusters and you can personalize feedback to certain cluster group. You know there are some students who might be spending much time in reading materials and taking quiz instead of watching any videos. For those learners, you might give feedback saying that A, why cannot you go and take look at the videos. So, those kind of feedback can go to those users when they were not able to give the answers to their quiz questions. For some learners who are not really interested in reading or doing any simulation, they just watch videos and try to answer the questions. For them, you can say why cannot you try simulators or some read materials. Or some learners who do not care anything, just directly go to assignment questions, try to answer them and complete the weeks assignment and go for next week. For them, you can give a different kind of feedback. So, you can provide a feedback if you cluster the students based on their behavior, the interaction with the system. Also, we can use this clustering to provide adaptive content based on the students interaction, you might give a hint or a new video or new course material, that is also possible. And there is other application of clustering, it is called e-learning materials. There are a lot of e-learning content available. We can go and crawl all the e-learning content and curate it. And you can automatically classify them into topic-wise or concept-wise. So, if you have those things, the system can automatically pick the e-learning content based on the topic you are interested or concept you are interested. So, that is also possible. So, clustering can be used to cluster the students behavior based on the interaction with the system or you can also content can be clustered or other things can be clustered in the system. So, in this video, we saw what is clustering and why it is needed for learning environments. In our next video, we will talk about one type of clustering. I will go through them and we will discuss in detail. Thank you.