Browse Prior Art Database

Aggregating Profiles For Complex Event Processing Based On Predictive Analytics

IP.com Disclosure Number: IPCOM000237955D
Publication Date: 2014-Jul-23
Document File: 4 page(s) / 89K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method and system is disclosed for aggregating one or more profiles for complex event processing based on one or more predictive analytics models for providing one or more recommendations.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 43% of the total text.

Page 01 of 4

Aggregating Profiles For Complex Event Processing Based On Predictive Analytics

A method and system is disclosed herein for aggregating one or more profiles using complex event processing based on one or more predictive analytics models for providing one or more recommendations. The method and system identifies the one or more profiles based on one or more metrics/measurements calculated from an event and retrieves the one or more profiles from a cache. The method and system, then, updates at least one profile of the one or more profiles by performing aggregations for the at least one profile. Thereafter, the method and system sends the one or more profiles retrieved from the cache to one or more scoring services (predictive models) and decision models in order to provide the one or more recommendations.

Fig. 1 illustrates the steps executed by the method and system for aggregating the one or more profiles for the complex event processing, according to embodiments of the present invention.

Figure 1

In accordance with fig. 1, the method and system receives an event as at least one of a single stream and a continuous stream. The method and system extracts one or more metrics from the event and optionally stores the event in a database. The method and system, then, feeds the event and profile data into the one or more predictive analytics models in order to provide one or more recommendations based on the complex event processing.

The disclosed method and system utilizes a combination of techniques such as, but not limited to, profile caching, batched profile selection and batched profile updates in order to perform the complex event processing. The combination of techniques is used in conjunction with one or more platform optimizations such as surrogate key caching and business key caching. The combination of techniques utilized by the one or more predictive analytics models are described below.

Profile caching:

The method and system stores the one or more profiles within the cache. When a profile is to be retrieved, the method and system forms a unique key using a combination of one or more business keys of the one or more profiles. The method and system, then, utilizes the unique key to look up for a profile in the cache. The

1


Page 02 of 4

method and system also addresses further requests to the profile via the cache. Here, the cache is designed to be local to at least one of an event and a stream of events arriving on a same thread.

Batched profile updates:

The method and system firstly identifies one or more profiles in the cache from one or more metrics/measurements calculated from the event. Based on the type of the event, the method and system performs calculations and aggregations of the one or more metrics/measurements for updating the one or more profiles pertaining to one or more assets/resources/processes described in the event. For example, a simple calculation can be a 'Measurement Above Limit' calculation for calcul...