SYSTEM AND METHODS FOR ANALYSIS OF DATA

• US 20150242469A1
• Filed: 09/27/2013
• Published: 08/27/2015
• Est. Priority Date: 09/27/2012
• Status: Active Grant

• Current Assignee: Cornell University
• Original Assignee: Cornell University
• Inventors: Lipson, Hod, Chattopadhyay, Ishanu
• Granted Patent: US 10,275,500 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2465 Query processing support for facilitating data mining operations in structured databases
• CPC Class Codes: G06F 16/319 Inverted lists
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 2218/00 Aspects of pattern recognition specially adapted for signal processing

Abstract

Data processing including a universal metric to quantify and estimate the similarity and dissimilarity between data sets. Data streams are perfectly annihilated by a correct realization of their anti-streams. Any deviation of the collision product from a baseline, for example flat white noise, quantifies statistical dissimilarity. The invention relates generally to data mining. More specifically, the invention relates to the analysis of data using a universal metric to quantify and estimate the similarity and dissimilarity between sets of data.

Claims

1. A computer method for analyzing data, comprising the steps of:

encoding a first data set to obtain a first encoded data set;

encoding a second data set to obtain a second encoded data set;

inverting the second encoded data set to obtain an inverted data set;

performing summation of the first encoded data set and the inverted data set to generate a summed data set;

encoding a baseline data set to obtain a baseline encoded data set;

comparing the summed data set to the baseline encoded data set; and

identifying one or more dissimilarities between the first data set and the second data set.

2. The computer method for analyzing data according to claim 1, wherein the baseline data set is flat white noise.

3. The computer method for analyzing data according to claim 1, wherein the first data set in encoded into a series of symbols.

4. The computer method for analyzing data according to claim 1, wherein the second data set in encoded into a series of symbols.

5. The computer method for analyzing data according to claim 1, wherein the baseline data set in encoded into a series of symbols.

6. The computer method for analyzing data according to claim 3, wherein the ordered series of symbols comprises a number 1 and a number 0.

7. The computer method for analyzing data according to claim 4, wherein the ordered series of symbols comprises a number 1 and a number 0.

8. The computer method for analyzing data according to claim 5, wherein the ordered series of symbols comprises a number 1 and a number 0.

9. The computer method for analyzing data according to claim 1, wherein the step of encoding a first data set further comprises the steps of:

quantizing the first data set to obtain a quantized data set; and

mapping one or more portions of the quantized data set to a symbol.

10. The computer method for analyzing data according to claim 1, wherein the step of encoding a second data set further comprises the steps of:

quantizing the second data set to obtain a quantized data set; and

mapping one or more portions of the quantized data set to a symbol.

11. The computer method for analyzing data according to claim 1, wherein the step of encoding a baseline data set further comprises the steps of:

quantizing the baseline data set to obtain a quantized data set; and

mapping one or more portions of the quantized data set to a symbol.

12. Algorithmic components of a computer method for analyzing data, comprising the steps of:

generating a first sample path from a hidden stochastic source;

generating a second sample path from the inverse model of the hidden stochastic source;

generating a third sample path from a sum of hidden stochastic sources;

estimating a deviation of a symbolic stream from flat white noise.