Artificial intelligence analyzer and generator

• US 7,512,570 B2
• Filed: 05/30/2006
• Issued: 03/31/2009
• Est. Priority Date: 05/30/2006
• Status: Expired due to Fees

• Current Assignee: Zaracom Technologies Inc.
• Original Assignee: Zaracom Technologies Inc.
• Inventors: Chang, Yuan-Lung, Chang, Yuan-Huei
• Primary Examiner(s): Holmes; Michael B
• Application Number:
  US11/420,879
• Publication Number:
  US 20070282768A1
• Time in Patent Office: 1,036 Days
• Field of Search: None
• US Class Current: 706/11
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

This invention relates to using artificial intelligence for analyzing real-life collected data from an operation system, modeling the collected data to identify characteristics of events, analyzing the models to conclude an optimal solution for maximizing the performance of the operation system.

Claims

1. A method of network analysis for a wireless network comprising,

providing a data partitioner to receive a first set of plurality types of data, wherein the data partitioner partitions the first set of data and groups the partitioned data into a plurality of data files, and the data partitioner generates indices according to the first set of data;

and

providing a data modeler to receive a second set of plurality types of data, wherein the data modeler generates a plurality of models according to the first set of data and the second set of data.

2. The method of network analysis for a wireless network of claim 1 comprising,

providing a model analyzer to receive a third set of plurality types of data wherein the model analyzer analyzes the plurality of models according to the third set of plurality types of data.

3. The method of network analysis for a wireless network of claim 2, wherein the indices indicate the data files and the partitioned data in the data files.

4. The method of network analysis for a wireless network of claim 3, wherein the indices are generated by using indexing algorithms comprising of R-Tree Indexing, Hash Code Indexing, Reverse Indexing, and Suffix Indexing.

5. The method of network analysis for a wireless network planning of claim 2, wherein the data partitioner partitions the first set of data by selecting an Average value of the first set of data, and by selecting data from eighty percent of normal distribution of the first set of data, and by selecting data from twenty percent of the first set of data wherein the twenty percent of the first set of data were not selected by the selection of the eighty percent of the normal distribution of the first set of data, and by selecting data of maximum values and minimum values from the first set of data, and by selecting data from common occurrence data of the first set of data, and by selecting data from non-common occurrence data of the first set of data.

6. The method of network analysis for a wireless network of claim 5, wherein the plurality of models are generated by Finite Impulse Response (FIR) algorithm, and Infinite Impulse Response (IIR) algorithm, and Segmentation algorithm, and Vectorization algorithm, and Fast Fourier Transform (FFT) algorithm, and Z-Transform algorithm, and Differential Run-length Huffman algorithm, and Down Sampling algorithm, and Geometrical Prediction algorithm, and Boundary algorithm, and Hierarchical Linear Modeling algorithm.

7. The method of network analysis for a wireless network of claim 6, wherein the model analyzer analyzes the plurality of models by using Singular Value Decomposition (SVD) algorithm, and Vertical Vectorization algorithm, and Vertical Grouping algorithm, and Horizontal Grouping algorithm, and Gravity-Velocity Analysis algorithm, and Bayesian Analysis algorithm, and Error Diffuse Iteration algorithm to generate an optimal model.

8. The method of network analysis for a wireless network of claim 7, wherein

the model analyzer generates Dynamic Constraints;

and

the model analyzer generates modified optimal models by new Dynamic Constraints.

9. The method of network analysis for a wireless network of claim 8, comprising:

an Optimizer, in accordance with the modified optimal models, generates an optimal system performance solution.

10. A method of network analysis for a wireless network, comprising:

providing a data partitioner to receive a first set of plurality types of data, wherein the data partitioner partitions the first set of data and groups the partitioned data into a plurality of data files, and the data partitioner generates indices according to the first set of data;

providing a model analyzer to receive a second set of data and generates optimal models and Dynamic Constraints.

11. The method of network analysis for a wireless network of claim 10, comprising,

providing a data modeler to receive a third set of plurality types of data, wherein the data modeler generates a plurality of models according to the first set of data and the third set of data.

12. The method of network analysis for a wireless network of claim 11, wherein the indices indicates data files and the partitioned data in the data files, and the indices are generated by using replicate tree indexing, hash code indexing, reverse indexing, and suffix indexing.

13. The method of network analysis for a wireless network of claim 11, wherein the data partitioner partitions the first set of data by selecting an Average of the first set of data, and by selecting data from eighty percent of normal distribution of the first set of data, and by selecting data from twenty percent of the first set of data wherein the twenty percent of the first set of data were not selected by the selection of the eighty percent of the normal distribution of the first set of data, and by selecting data of maximum values and minimum values from the first set of data, and by selecting data from common occurrence data of the first set of data, and by selecting data from non-common occurrence data of the first set of data.

14. The method of network analysis for a wireless network of claim 13, wherein the plurality of models are generated by Finite Impulse Response (FIR) algorithm, and Infinite Impulse Response (IIR) algorithm, and Segmentation algorithm, and Vectorization algorithm, and Fast Fourier Transform (FFT) algorithm, and Z-Transform algorithm, and Differential Run-length Huffman algorithm, and Down Sampling algorithm, and Geometrical Prediction algorithm, and Boundary algorithm, and Hierarchical Linear Modeling algorithm.

15. The method of network analysis for a wireless network of claim 14, wherein the model analyzer analyzes the plurality of models by using Singular Value Decomposition (SVD) algorithm, and Vertical Vectorization algorithm, and Vertical Grouping algorithm, and Horizontal Grouping algorithm, and Gravity-Velocity Analysis algorithm, and Bayesian Analysis algorithm, and Error Diffuse Iteration algorithm to generate optimal models.

16. The method of network analysis for a wireless network of claim 15, wherein the model analyzer generates optimal models by Dynamic Constraints;

and

the artificial intelligent system further comprises an Optimizer generates an optimal performance solution based on one of the optimal models.

17. A method of network analysis for a wireless network, comprising:

providing a data partitioner to receive a first set of plurality types of data, wherein the data partitioner partitions the first set of data and groups the partitioned data into plurality of data files, and the data partitioner generates indices according to the first set of data;

providing a data modeler receives second set of plurality types of data, wherein the data modeler generates a plurality of models according to the first set of data and the second set of data; and

providing a model analyzer to receive a third set of data wherein the model analyzer analyzes the plurality of models according to the third set of data.

18. The method of network analysis for a wireless network of claim 17, wherein the indices indicates data files and the partitioned data in the data files;

and

the indices are generated by using replicate tree indexing, hash code indexing, reverse indexing, and suffix indexing.

19. The method of network analysis for a wireless network of claim 18, wherein the data partitioner partitions the first set of data by selecting an Average of the first set of data, and by selecting data from eighty percent of normal distribution of the first set of data, and by selecting data from twenty percent of the first set of data wherein the twenty percent of the first set of data were not selected by the selection of the eighty percent of the normal distribution of the first set of data, and by selecting data of maximum values and minimum values from the first set of data, and by selecting data from common occurrence data of the first set of data, and by selecting data from non-common occurrence data of the first set of data;

and

the plurality of models are generated by Finite Impulse Response (FIR) algorithm, and Infinite Impulse Response (IIR) algorithm, and Segmentation algorithm, and Vectorization algorithm, and Fast Fourier Transform (FFT) algorithm, and Z-Transform algorithm, and Differential Run-length Huffman algorithm, and Down Sampling algorithm, and Geometrical Prediction algorithm, and Boundary algorithm, and Hierarchical Linear Modeling algorithm.

20. The method of network analysis for a wireless network of claim 19, wherein the model analyzer analyzes the plurality of models by using Singular Value Decomposition (SVD) algorithm, and Vertical Vectorization algorithm, and Vertical Grouping algorithm, and Horizontal Grouping algorithm, and Gravity-Velocity Analysis algorithm, and Bayesian Analysis algorithm, and Error Diffuse Iteration algorithm to generate Vendor Independent Models, and Pattern-Relationship;

and

the model analyzer generates optimal models and Dynamic Constraints; and

the artificial intelligent system further comprises an Optimizer wherein the Optimizer generates an optimal solution based on one of the optimal models.