The Power of Multivariate Analysis
Optimus RP is the only tree-based tool that can perform true multivariate analysis, providing the ability to look at multiple outcomes simultaneously. This allows you to find patterns you would not find in separate analyses of individual dependent variables.
While it may be possible at times to model each target separately, or perhaps create a single covariate as a proxy for several and model for it, multivariate analysis makes this unnecessary, and uncovers unanticipated as well as expected departures from the normal patterns of relationship – meaning that better predictions are possible using all outcomes simultaneously.
Examples
- Financial: Simultaneously predict risk of default coupled with the likelihood of early payoff, balanced against risk of fraud, or perhaps attrition balanced with profitability and cross-sell opportunities.
- Marketing: Model campaign effectiveness alongside the predicted lifetime value of a customer.
- Internet Search Providers: model click-through with predicted goal conversion rate.
- Pharmaceutical: Simultaneously model for absorption, distribution, metabolism, excretion, and toxicity (ADME-Tox) of a compound.
Without Optimus RP, analyses such as these would be difficult, less powerful, and require the creation of several, unrelated models. In other words, it would take longer to develop, require more maintenance, and be less powerful.
The example below demonstrates how multivariate approaches can find significance that other tools will miss. Notice that the difference between the two populations is not significant in x or y alone. However, in both x and y, the difference is highly significant
