 Welcome back to learning analytics course. So, in the last video you have performed a diagnostic analytics such as why a student did not pass the exam or why the student was not able to answer certain questions. You collected the data such as students performance also the attendance. Now I want you to think how will you predict the students performance in the future events. For example, you found a correlation that students attendance and performance are side correlation. Can you predict what will happen in future? Please pass this video, write down your answers after writing it down, resume the video to continue. So, what is predictor analytics? So, it is predicting what will happen next. So, which course will have a less number of registrations. This can be predictions because based on the last future historical data, we can predict what will happen in next year. Which students will not complete the course based on the students attendance or students other interaction behavior, you can predict whether the student can complete the course or not and this again using the model which you created using the historical data. What will be the performance of the student in the next question? So, you can go much finer level like if he is interacting with the intelligent learning environment or technology enhanced learning environment. What will be the students next set of actions? Will the student able to answer the next question correctly? All those final level analysis we can do. So, prediction is done based on data from past events. In the sense, if you are teaching a class in this year, you cannot predict, create a model using current data. Instead, you have to use the data from last 4-5 years car data. Like as I mentioned, you have last 5 years data. From that last historical data, you might be creating the model. Apply that model in this year. That is what it says that you have to create the model based on past events. Also in some cases, we will use the present data also because that will help to improve the algorithm to predict it better. It is most popular in educational data mining. It is called learner modeling. We try to model the learner and also it is popular in learning analytics. There are lot of tools available for teachers and other stakeholders to perform the predictive analytics. However, we need to understand what algorithm to apply and what data will suitable for it. And you have to understand how to interpret the results when the algorithm gives. So, we will talk about Naive Bayes decision tree in this class and the tools like Orange and Wakeup to use these algorithms on education data. So, in short, prior to analytics is extracts information from data sets in order to detain pattern and predict the future events. It uses both past and present data and offer prediction for the future. So, present data actually is an input variable to the model and those model predicts what will happen in the future given the present data. So, let us move on to the last activity of this work. Now that you are able to predict whether the learner will continue your course or drop it, assume that what measure will you take to ensure your learners are motivated enough to continue your course and out. For example, assume that you have created a correlation between attendance and performance and you know that learner correlation is high with attendance. Using that information you can predict whether the learner will drop the course or whether the learner will fail in the exams. If you have that prediction, what measures will you take in order to motivate the students and continue doing the course. Please pause this video, write down your answers. After writing it down, resume this video to continue. Prescript analytics is if you say that I want scaffold to students give hints of feedbacks to help students to achieve the learning goal that can be one of the prescript analytics. Or you can use a personalized or intelligent learning environments where based on the students interaction and performance, the system gives a feedbacks and hints and adapts the content. Or you can predict the learner's current state and provide feedback and help them achieve the learning goal. So, you can have special classes and you can talk to them what is the problem. It is not that a student will fail or student will drop out. You want to understand why also why the student will not able to complete the course. If you know why the reasons then you will be able to provide a proper informed feedback or adaptive content. So, the prescript analytics is a process of analyzing the data and providing instant recommendation to how to optimize the learner's learning process. For example, an instructor can use a prescript analytics to discover that most of the learners needed a prerequisite course before joining a newly launched advanced course. So, based on the previous experience or based on the teaching experience, you can so say that for the newly launched advanced course given the learners entry exam, you might need a prerequisite course. So, the teacher might conduct a test before learner taking the course. Based on the student's performance, a teacher can decide whether the learner needs a prerequisite or not. So, that is based on the teaching experience or the predictive model you are created. I want to inform you that we saw the four types of analytics in last couple of videos. So, if we talk about prescriptive, it subsumes predictive, diagnostic and descriptive analytics. Which means if I doing a predictive analytics, I should be doing the descriptive and diagnostic analytics also. It is not that I can pick and do only the predictive analytics, I can pick only do the prescript analytics. So, you can start always with the descriptive analytics, then do the diagnostic analytics, then go for the predictive analytics. In this course, we will cover the three types of analytics like descriptive, diagnostic and predictive. Prescript analytics is beyond the scope of this course because prescript analytics is designing a intelligent tutoring systems. So, in this video, we described what is predictive analytics and what is prescriptive analytics and you might have had some idea on that. We will in detail discuss about these two type of analytics in our course. Thank you.