Week 4: Predictive Models & Assessment

Week 4: Predictive Models and Assessment

Introduction:

A key focus of many analytics projects is prediction. With a sufficient quality sample of data, patterns can be discovered that can then lead to predictions based on other learners that share those patterns. These predictions, whether marketing, learner succes, or healthcare needs, can be surprisingly accurate and provide individuals and organizations with valuable guidance on resource allocation or needed interventions.

As you'll note in the readings and videos this week, many predictive models are adapted from healthcare and business. A few projects (Purdue Signals, Rio Salado's predictive model) are evident, but not the norm. It's not that predictive models are difficult. The difficulty is that any system that includes autonomous agents will continue to evolve. In a mechanistic system, prediction is easier because variables are consistent. Social systems consist of people doing unpredictable things. As Yogi Bera stated: "It's difficult to make predictions. Especially about the future".

Readings and Videos:

Learning Activities:

1. Upload your Analytics: Logic and Structure to the Post Link to (or attach) Assign 1

2. Participate in Week 4 Discussion Forum. The secondary theme of this week is on assessment. After reviewing the resources above, engage in the discussion forum on how predictive models can contribute to improving assessment. LA and predictive modeling has significant potential to advance the work psychometricians. What are some specific areas of impact?

3. Continue working on your Concept Map

4. Continue working Analytics Project