What is it ?
Prescriptive Analytics lets you simulate and optimize your business.
The goal of prescriptive analytics is to be able to tell you what action has the highest likelihood of a beneficial effect when a predicted event does occur.
Prescriptive analytics require strong predictive analytics as a foundation for how the future will look ceteris paribus.
What prescriptive analytics does is to take the predicted future and let you affect it. It lets you alter the different input variables in the predictive models and see how it affects the predictions.
In other words it lets the user simulate the future. Good prescriptive analytics is bound by ranges determined by the predictive analytics. That will let you simulate within significant ranges, which will help improve model precision.
Prescriptive analytics is not only simulation but also optimizing. It will run optimization algorithms to automatically make suggestions into which inputs in the simulations that is most beneficial for your case.
You can of course optimize on different parameters. For example, it makes a
difference whether you are risk averse or not, which will be reflected in the
Here is some inspiration
Prescriptive analytics is about modelling on data and dependencies discovered in earlier stages of analytics.
- Analytical Financial Statements
- Product Pricing
It takes the simulation options and puts them in a framework where you optimize on pre defined conditions. Examples:
Optimize with regards to
- Highest profit/value
- Lowest risk
- Weights from inputs
- Lowest model uncertainty
Prescriptive Analytics has high demands.
Prescriptive analytics is the third and last leg of the three analytics that Ivin does.
It is highly correlated with predictive analytics as it lets you simulate and optimize on your predictions from earlier analytics stages. It thereby requires access to all of the intermediate steps from the predictive analytics.
Partly as a benchmark against the model precision and mostly because it will often use the output, or ranges of the output, from the predictive analytics as boundaries and assumptions in the simulations.
Ivin does not require that the foundational predictive analytics is implemented by Ivin, as long as it is well documented.
Just like Ivin has no problem with other companies building prescriptive analytics on predictive analytics created by Ivin.