To be an informed consumer of predictive modeling you need to be aware of the modeling tools that will be used to create your particular model. Many practitioners are trained and experienced in only one tool and tend to use it in all cases. Not all tools are equally appropriate or accurate in all cases.

PreMo is tool agnostic

We constantly strive to keep abreast of the evolving availability of new techniques and packages and our senior model developers will as a group be qualified in essentially all of the most current techniques including:

    Artificial Neural Nets
    Genetic Programs
    Support Vector Machines
    Regression (MARS)
    Bayesian Classifiers
    Decision Trees (CART)
    OLAP Analysis and Data Visualization

We also have professional associates highly qualified in SAS and SPSS if this is a requirement.

Model Accuracy

Clients are frequently so thrilled to have the benefits of a good model that they do not stop to ask if this is the best model. The accuracy of your model will depend on three things: 1). First and foremost the ability of your data to be predictive. There is an unknown and fixed limit to which any data can be predictive regardless of the tools used or experience of the modeler. 2.) The experience and skill of the senior predictive modeler assigned to you. Their experience with similar problems, with the nature of your data, and the tools available will be a major factor. 3.) The tools selected. Some tools are designed to give very quick if somewhat approximate results. Other tools are inherently more accurate if somewhat slower. PreMo always uses at least two leading tools or techniques to create your model to ensure the answer you get will be the most accurate possible.

How Important is Model Accuracy?

Read our case study on how a .01 change in accuracy created an 8% improvement in financial outcome. Small changes in accuracy can mean big improvements in your bottom line.

Our Favorite Tool

OK we said we were tool agnostic and we are. But just as in any field where many tools are available there is one that we always reach into our toolbox for whenever we can, the Discipulus Genetic Program. We have run a series of tests between Discipulus and other leading tools (neural nets, support vector machines, logistic regression, Bayesian classifiers) and Discipulus has never lost (occasionally tied indicating that both tools has extracted the maximum predictive information from the data). Read more about our favorite tool and why it can deliver this extraordinary accuracy.

Are You Already a User of Predictive Models?

Our Challenge to You

If you are currently using predictive modeling PreMo will provide a no cost/no risk assessment of the accuracy of your current models. We believe that the combination of the tools we use and the experience of our practitioners can improve the accuracy of most existing models. Greater accuracy means increased revenues, greater profits, and decreased risk. [Details]