Predictive Analytics

Catalog description

In most business situations, being able to determine the value of some unknown with reasonable accuracy can be beneficial. For example, it would be useful for a company to know the if prospective customer would default on payments (classification), or to know the number of units of a product that it might be able to sell during the next quarter at a given store (regression). Quite often, even seemingly inaccurate estimates of such unknowns can lead to large monetary gains for a company if the new knowledge can lead to a discernable difference in performance. This is the domain of predictive modeling – using historical data to determine the value of an unknown. The course covers both classification and regression techniques. Of course, just any model will not do; analysts need to ensure that their models are valid and will work on new data and the course covers approaches to model validation. The course will equip participants with the ability to identify situations that could benefit from predictive models, to identify the data requirements and work with concerned people to get the data, to manipulate the data into a form usable for predictive analysis, and to build, evaluate, present and deploy the models.