SYSTEM AND METHOD FOR CONTEXT MAP GENERATION

• US 20100011021A1
• Filed: 07/11/2008
• Published: 01/14/2010
• Est. Priority Date: 07/11/2008
• Status: Active Grant

• Current Assignee: Satyam Computer Services Limited of Mayfair Centre
• Original Assignee: Satyam Computer Services Limited of Mayfair Centre
• Inventors: Kumar, G.S. Praveen, Kurup, Swetha Mukunda, Varadarajan, Sridhar, Gopalan, Srividya
• Granted Patent: US 8,538,958 B2
• Time in Patent Office: Days
• US Class Current: N/A
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

Personalized services and targeted services demand the exploitation of context—say, context of an environment, context of a product, context of a service, and context of a customer or user or individual. In order for the personalized and targeted services to have a greater impact, the identified context used by a context aware processing engine needs to be accurate and current. Such an up-to-date context is called as context map and a system and method for generating a customer-specific context map involves the processing of customer transactions, abstracting and generalizing these transactions into entities, and inter-relating these entities. The context map, for example, is useful in ad targeting in mobile environments, and in quicker and efficient problem resolution in contact center environments.

Claims

1. A method for the creation of a context map with respect to a customer, the said method comprising:

receiving of a plurality of transactions associated with said customer, wherein said plurality of transactions include a plurality of voice calls made and received by said customer, a plurality of text messages sent and received by said customer, a plurality of web pages browsed by said customer, a plurality of emails sent and received by said customer, a plurality of images captured by said customer, a calendar information as part of personal information updated by said customer, a plurality of products bought by said customer, a plurality of services subscribed to by said customer, and a plurality of enquiries made to a contact center regarding said plurality of products and said plurality of services, receiving of a database of customer ontology, wherein said database of customer ontology includes a plurality of domain specific entities, a plurality of relationships among said plurality of domain specific entities, and an annotation of each of said plurality of relationships, wherein said annotation of a relationship of said plurality of relationships includes a plurality of correlation descriptions among a plurality of entities associated with said relationship, analysis of said plurality of transactions with respect to said database of customer ontology to determine a database of entities with respect to said customer, and analysis of said plurality of transactions associated with said customer and said database of entities with respect to said customer to determine said context map with respect to said customer.

2. The method of claim 1, wherein said method further comprising:

analyzing voice data associated with a transaction of said plurality of transactions associated with said customer using an automatic speech recognition technique to extract a plurality of keywords, mapping each of said plurality of keywords to an entity of said database of customer ontology or an attribute of an entity of said database of customer ontology, and incrementing count of said entity or said attribute of said entity in a database of entity counts.

3. The method of claim 2, wherein said method further comprising:

analyzing text data associated with a transaction of said plurality of transactions associated with said customer using a text analysis technique to extract a plurality of keywords, mapping each of said plurality of keywords to an entity of said database of customer ontology or an attribute of an entity of said database of customer ontology, and incrementing count of said entity or said attribute of said entity in said database of entity counts.

4. The method of claim 2, wherein said method further comprising:

analyzing image data associated with a transaction of said plurality of transactions associated with said customer using an image analysis technique to extract a plurality of keywords, mapping each of said plurality of keywords to an entity of said database of customer ontology or an attribute of an entity of said database of customer ontology, and incrementing count of said entity or said attribute of said entity in said database of entity counts.

5. The method of claim 2, wherein said method further comprising:

selection of an entity from said database of entity counts if count associated with said entity in said database of entity counts exceeds a pre-defined threshold, selection of an entity with a plurality of attributes from said database of entity counts if count associated with each of said plurality of attributes in said database of entity counts exceeds a pre-defined threshold, and making of selected entity a part of said database of entities with respect to said customer.

6. The method of claim 1, wherein said method further comprising:

determination of a plurality of ECO entities based on said database of customer ontology, determination of frequency of an entity (Mi) of said plurality of ECO entities based on said database of entities with respect to said customer and said plurality of transactions associated with said customer, computation of sum (M) of Mi with i varying from 1 to N, wherein N is the number of said plurality of ECO entities, selection of an entity Ei from said plurality of ECO entities if Mi/M associated with said entity Ei exceeds a pre-defined threshold, and making of said entity Ei a part of a database of selected entities.

7. The method of claim 6, wherein said method further comprising:

identification of a relation R based on said database of customer ontology, identification of a plurality of entities with respect to said relation R, identification of a plurality of attributes with respect to each of said plurality of entities, generation of a database table Tr with as many attributes of as many entities instantiated and normalized as possible based on said database selected entities, identification of a plurality of correlations Rc={CD1, CD2, . . . , CDm} associated with said relation R, evaluation of each of said plurality of correlations CDi based on Tr, said database of entities, and said plurality of transactions associated with said customer to determine a pair of correlation coefficients ACi and BCi, wherein ACi is an absolute measure based on total tuples in Tr satisfying CDi and total tuples with respect to CDi in Tr, and BCi is a relative measure based on ACi and sum of ACi with i varying from 1 to M, wherein M is the number of correlations in said plurality of correlations Rc, and making an instance of CDi a part of said context map if ACi exceeds a pre-defined threshold and BCi exceeds a pre-defined threshold.