Quebec Exploration

Technology

MCubiX

The Corporation designed and developed MCubiX, a data mining software suite capable of extracting knowledge from historical data - in the form of images, text or traditional databases - and making predictions that improve business processes and decision making. Our data mining solutions include:

  • MCubiX-KE (Knowledge Extractor)
  • MCubiX-EP (Entreprise Predictor)
  • CARDS (Computer Aided Resources Detection System)

MCubiX-KE

MCubiX-KE is a complete data mining toolbox that includes a wide selection of algorithms (decision trees, neural networks, association rules, etc.) and visualisation tools to help bring solutions to the most complex problems. Rules discovered with MCubiX-KE can be exported in PMML and XML in order to instantly feed MCubiX-EP scoring engines or to simplify hard coding the rules in a host application. The intuitive interface of MCubiX-KE is easy to use and guides the data miner through all modeling steps without the need for specialised programming knowledge. The result? Powerful models created in few clicks giving you the ability to make informed business decisions.

MCubiX-EP

MCubiX-EP is a family of scoring engine used to implement the rules discovered with MCubiX-KE within host applications of all kinds. MCubiX-EP allows scoring of local data by using the Windows service or of network data via the WEB service.

MCubiX-EP was developed with the end-user in mind. End-user applications can now take full advantage of data mining technologies. Using our MCubiX-EP, data mining can be embedded directly into any end-user application thus isolating the end-user from the complexities of modeling and scoring of a database. Data miners develop and fit models with MCubiX-KE and store their models in PMML and/or XML format in a central repository. This repository is in turn consulted by the MCubiX-EP modules to obtain the rules established by the data miner in order to score the dataset(s) presented by the host application.

MCubiX-EP consists of the following modules:

  • an API (MCubiX-API) that allows developers to embed our scoring engine into their host applications;
  • a Scoring Engine (MCubiX-SE) that allows scoring large datasets using;
  • a Web Service that allows scoring on servers reached via the Internet;
  • a Windows Service that allows scoring over a LAN or Intranet (Windows based). For more technical information, please download the MCubiX-EP technical brochure.

CARDS

CARDS is a vertical application developed by DIAGNOS to assist scientists and professionals of natural resources exploration companies in their search for profitable sites. CARDS uses MCubiX-KE’s, the latest artificial intelligence and pattern recognition algorithms to learn the “signatures” or “finger prints” of resources areas with a high economic potential and to automatically detect similar sites across unexploited regions. The result? A powerful tool that allows prospectors to spend more time exploring high potential areas.

CARDS' use is simple. CARDS requires two types of data sets: a training set, used to discover the signature or fingerprints of highly profitable regions (compared to unprofitable ones), and one or many surveys describing the territory that is prospected. The final output is a map or layer that indicates the regions where economical resources are likely to be found.

CARDS can be applied to prospecting for diamonds, gold, copper, zinc, copper, oil and gas, etc. Each resource requires a special methodology supported by CARDS.