Method and apparatus for interpreting information

• US 20040093316A1
• Filed: 10/31/2003
• Published: 05/13/2004
• Est. Priority Date: 04/15/1997
• Status: Active Grant

• Current Assignee: Cerebrus Solutions Ltd (Neural Technologies Limited)
• Original Assignee: Cerebrus Solutions Ltd (Neural Technologies Limited)
• Inventors: Field, Simon, Howard, Gary, Hobson, Philip William, Barson, Paul Colin
• Granted Patent: US 7,275,047 B2
• Time in Patent Office: Days
• US Class Current: 706/25
• CPC Class Codes: G06F 18/24765 Rule-based classification
• CPC Class Codes: G06N 3/042 Knowledge-based neural networks; Logical representations of neural networks
• CPC Class Codes: G06N 5/025 Extracting rules from data
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

A method of processing data relating to a plurality of examples using a data classifier arranged to classify input data into one of a number of classes, and a rule inducer, comprising the steps of:

(i) inputting a series of inputs to the data classifier so as to obtain a series of corresponding outputs;

(ii) inputting said series of outputs and at least some of said series of inputs to the rule inducer so as to obtain a series of rules which describe relationships between the series of inputs to the data classifier and the series of corresponding outputs from the data classifier.

Claims

1. A method, comprising:

generating classification outputs from a first data classifier based on first data elements,

using the classification outputs and the first data elements to generate rules relating the first data elements and the classification outputs.

2. The method of claim 1, wherein the first data elements include data based on usage of a telecommunications network.

3. The method of claim 2, wherein the data include call detail records.

4. The method of claim 1, wherein the first data classifier includes at least one of:

a neural network, an unsupervised classifier, and a Bayesian classifier.

5. The method of claim 1, wherein generating rules includes:

combining the first data elements and the corresponding classification outputs.

6. The method of claim 1, further comprising:

based on the generated rules, generating a second data classifier.

7. A system, comprising:

a first data classifier configured to generate, based on first data elements, classification outputs, and

a rule inducer configured to use the first data elements and the classification outputs to generate rules relating the first data elements and the classification outputs.

8. The system of claim 7, wherein the first data elements include data based on usage of a telecommunications network.

9. The system of claim 8, wherein the data include call detail records.

10. The system of claim 7, wherein the data classifier includes at least one of:

a neural network, an unsupervised classifier, and a Bayesian classifier.

11. The system of claim 7, further comprising:

a second data classifier based on the rules.

12. A processor program for classifying data, the processor program disposed on a processor-readable medium and comprising instructions to cause a processor to:

generate classification outputs based on first data elements provided to a first data classifier, and

generate rules relating the first data elements and the classification outputs, based on the first data elements and the classification outputs.

13. The processor program of claim 12, wherein the first data elements include data based on usage of a telecommunications network.

14. The processor program of claim 12, further comprising instructions to cause to a processor to:

generate a second data classifier based on the rules.

15. The processor program of claim 12, where the first data classifier is at least one of:

a neural network, an unsupervised classifier, and a Bayesian classifier.