• financial analysts wanting more accurate risk-based pricing models
• professional traders looking for superior technical trading signals
• database marketers in need of better response and cross-selling models
• internet retailers desiring to improve their product recommendations
• researchers seeking to develop intelligent systems from large corpora
• software developers seeking to add intelligence to their applications

"Using data supplied by Standard & Poor's, Mnemonic developed a prototype mortgage scoring model that is more robust than any reviewed by Standard & Poor's to date. When its accuracy is compared to that of several other mortgage scoring models, including the current Standard & Poor's model, the Mnemonic model significantly outperforms the others." -- Standard & Poors

The PMT provides proven predictive modeling techniques which are not available from any other commercial software vendor. It can be used by itself as your only data mining package or to complement your existing data mining software.

The great strengths of the PMT are:

• Easy to use: The modeler defines the variables and prepares the data; the PMT automatically induces models from the training data, and validates them on the testing data. No need to waste your time extracting features or fiddling with dummy variables.
• Masters "subtle effects": The PMT allows modelers to accurately model subtle dependencies among variables in a way that is not possible with traditional data mining techniques.
• Unmatched statistical efficiency: The PMT allows modelers to approximate complex high order conditional distributions with reliable low order statistics. It lets you model the intersection of complex events without fragmenting the training data or overfitting.
• Eliminates unwarranted assumptions: The PMT learns exactly what the data says, no more and no less, unlike software packages based on decision trees, logistic regression, and neural networks, which may make spurious assumptions about their data.
• Uses all available data: Model accuracy depends on the amount of training data. Unlike other statistical packages, the PMT does not limit the size of your data set. The PMT has induced models from billions of records.