XELOPES

from Wikipedia, the free encyclopedia

XELOPES is a real-time analytics library that contains numerous algorithms from the areas of data mining and reinforcement learning and is used for learning and decision-making in real time. Typical applications of XELOPES are recommendation engines , systems for real-time scoring and dynamic disposition . The XELOPES library has both a commercial and an open source version.

XELOPES has been jointly operated by prudsys AG and ZSoft Ltd. since 2001 . and the Electrotechnical University of Saint Petersburg .

concept

XELOPES is essentially based on the Common Warehouse Model (CWM) of OMG for handling metadata in data warehouses and primarily extends the data mining package of the CWM. The focus here was on adding real-time features such as data streams , adaptive learning methods and agents .

Mathematically, the XELOPES library is based on a continuous modeling of the analysis processes using operator equations , the solution of which is described using basic transformations . The mathematical framework is very abstract, but allows a strong standardization of different analysis methods.

properties

The essential properties of the XELOPES are:

  • platform-independent modeling in UML
  • Implementations in C ++ , Java and C #
  • Packages for data access, transformation, OLAP , statistics, data mining, text mining , reinforcement learning, dynamic disposition and price optimization
  • Comprehensive data stream concept for real-time processing of data streams
  • uniform mathematical modeling of all processes as basic transformations
  • numerous newly developed analysis methods such as thin mesh regression, nonlinear decision trees and hierarchical reinforcement learning
  • Serialization of all analysis models via extended PMML format

proof

Web links