 |
Psydex provides seamless real time analytics over both streaming and historical data sets bridging the gap between Complex Event Processing and traditional Business Intelligence Systems.
Signals and alerts are generated in real time from statistics and correlations of Semantic Topics and Data Topics. The Psydex Rules Engine processes inputs such as Topic model counts, statistics, co-occuring Topics and Time series to generate high-order events with extreme speed and accuracy
Leveraging massive parallelization, Psydex AG enables Terabytes of Real Time Streaming and Historical data to be seamlessly queried (ad hoc) in a fraction of a second to deliver Time Series, Statistical values and Correlations for complex PQL expressions involving thousands of words, phrases and dimensions.
Statistics represent source weighted frequency counts for models across varying period/intervals combinations (e.g. 1 day/1 minute, 30 day/daily). Every Word Token that moves through the Analytics Grid causes statistical updates for all matching models. Statistics are based on custom source weightings that can be defined by users. Topic Models are always processed in an Ad Hoc fashion enabling Models to evolve indepedently
|
Topic Statistics are updated in real time.... Models can change in Real Time. |
|
Correlation coefficents for Topics are generated in real time against external data sets such as stocks and commodities
The Psydex Event Processor System leverages Topic Statistics and Time Series to more accurately understand the relevancy of current Events and Topics based on historical patterns and trends
|
|
