|Building synthetic brains capable of human level discovery and invention...|
ClassAct allows the user to automatically cluster database entries into self-similar families with the utmost speed and efficiency over other sorting/classification algorithms. Therefore, given the taxonomic problem of sorting animal species into similar classes, from a database of animals and their attributes (i.e., Boolean values such as has bony spine, has hair, give milk, etc.), ClassAct automatically separates the database into natural classes, perhaps tantamount to known species, or a more efficient scheme than present day biology has cumulatively arrived at. ClassAct may also be used in the anomaly detection mode to determine which, if any, of your database entries are more anomalous than others. Used in this manner, this application can easily comb your database for outliers, perhaps the result of bad measurements, or typographical errors.
No complicated setup. Just point the application at the database and press the toolbar button to cluster your records.
Application automatically senses the dimensionality of your problem and appropriately adjusts the neural architecture.
Slider control adjusts the granularity of classification.
Classify records having thousands of numerical attribute fields.
Export your clustered database for further use or study.
ClassAct uses a Creativity Machine to best decide how to arrange data patterns into self-similar groups, essentially minimizing intra-class Euclidean distance between similar patterns and optimizing inter-class distance between such classes. In effect, such neural nets work just like our brains, essentially categorizing things and scenarios within the external environment, before deciding upon a course of action. But recall, that IEI does not use ordinary artificial neural networks. Instead it uses IEI's own, ultra-fast, ultra-efficient neural network called a STANNO. Accordingly, the classification layers of these networks work at astounding speeds, allowing very high dimensional patterns within a database to be quickly aggregated into their natural families even on humble desktop PCs, within seconds.
To use ClassAct, simply open your database from the File menu, then press the Classify tool bar button. Within seconds, you will see your database records segregating into its natural families. To determine outliers in your database, press the tool bar button denoted by thed.
Using ClassAct, the user could, for instance:
Aggregate a client database into customers having similar demographic attributes. Having located your best customer, for instance, clients within his/her class will be most likely to buy your product or service.
Cluster employees together into self-similar groupings to identify those with similar talents, while likewise identifying any potential vocational 'gaps' within your organization.
In a law enforcement capacity, self-organize modus operandi databases into their natural classes to focus the search for prime suspects.
Quickly comb neural network training exemplars to identify outliers prior to training using products like IEI's PatternMaster.
© 2011, Imagination Engines, Inc.
All Rights Reserved
|Imagination Engines, Inc., 1550 Wall Street, Ste. 300, St. Charles, MO 63303, (636) 724-9000|
|For technical support, contact firstname.lastname@example.org.|
|For contracts, contact email@example.com.|