About Us

Pattern Discovery Technologies Inc. is a pioneer in the field of data mining and predictive analytics.

Founded in 1997 as a spin-off from the world renowned Pattern Analysis and Machine Intelligence (PAMI) Lab at the University of Waterloo, Pattern Discovery continues to push the envelope in developing solutions that tackle the most challenging data mining, analysis and optimization problems for industrial processes.

Our products and services are used by leading companies engaged in complex industrial processes such as petroleum refining, mineral and metal extraction, energy and renewables, food processing, and pharmaceutical operations, to name a few.

Leveraging their existing investments in process instrumentation, these companies search for clues hidden in the volumes of data generated daily that will uncover opportunities for continuous improvement in their operations.

We are also actively expanding our business into new industries, including mineral & metal production, power generation, and water treatment operations.

The ability to characterize key performance factors and forecast events using Intelligent Analytics can return crucial dividends. In operations that are complex and multi-faceted, count on Pattern Discovery’s state-of-the-art technologies to provide the insight needed to take control of the situation.

• Discover new & unique relationships that are not intuitively obvious
• Optimize processes and resources by understanding and quantifying the key parameters that drive operational efficiencies
• Predict results that take into account the interrelationships of complex multi-dimensional factors (forecast events before they happen)
• Make decisions that can be acted on more confidently and more consistently, using a basis that is squarely supported by the data that drives the operation (free of assumptions and personal biases)
• Leverage existing investments in data collection and aggregation by integrating Discover*e to provide advanced analysis solutions

Predicting with confidence is a powerful tool in any decision making process. To facilitate confident prediction, the descriptive patterns found are then transformed into production rules for predictions. The strength of a rule is measured by Weight of Evidence (WoE), which provides evidence that a specific pattern contributes to a target event. Following this same approach, new data can be classified using the rules established, and the WoE of each rule is accumulative in support of the classification with partial information.

This approach provides a totally unbiased, data driven approach to data mining and predictive analytics. It works with high order relationships, taking into account all of the factors that should be considered in performing analysis of complex datasets. The results of both the data mining exercise and the prediction are descriptive in nature meaning they can be interpreted in plain English by subject matter experts to derive greater insight into their process or investigation. And knowing the reasons why a particular prediction is being made can lead to the establishment of a proper set of actions or decisions based on the predicted result.