# US 2019/0171833 A1 · Protection of data privacy for cognitive demand forecasting

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0171833 A1
- Document ID: US20190171833A1
- Status: Published application
- Kind: A1
- Application number: 15/831,418
- Filed: 2017-12-05
- Published: 2019-06-06
- Earliest claimed: 2017-12-05
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Xu Nan Zhang; Ming Chao Wan; Miao He; Lin Tang; Yue Tong; Bing Shao; Changrui Ren
- Web page: https://patentgrep.com/patent/US20190171833A1

## CPC

- G06F 21/6218: {to a system of files or objects, e.g. local or distributed file system or database}
- G06F 21/6254: {by anonymising data, e.g. decorrelating personal data from the owner's identification}
- H04L 9/0618: {Block ciphers, i.e. encrypting groups of characters of a plain text message using fixed encryption transformation}
- H04L 63/0428: {wherein the data content is protected, e.g. by encrypting or encapsulating the payload}
- G06Q 30/0202: Market predictions or forecasting for commercial activities
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

Machine logic (for example, software) for protecting sensitive and confidential user data and sending the protected user data to an authorized third party for predictive analytics purposes. The machine logic rules perform linear transformation operations on the sensitive and confidential user data in order to obscure the underlying data before it is sent to an authorized third party for performing predictive analysis of the underlying data. After performing predictive analysis on the obscured data by the authorized third party, the owner of the sensitive and confidential data takes a consequential action (that is, the owner of the data makes a business-related decision) based on the prediction.

## Claims

### Claim 1

A method comprising:
receiving a first time-series data set, with the first time-series data set including: (i) a plurality of time-series values, and (ii) a relative order of the first plurality of time-series values;
performing a linear transformation on the first time-series data set to obtain a first transformed time series data set, with the linear transformation being a set of mathematical operation(s) that: (i) change each value of the first plurality of time-series values respectively corresponding transformed values, and (ii) maintains, in the transformed values, proportions that exist among and between the values of the first plurality of time-series values;
sending, over a communication network and to a remote computer device, the transformed time-series data set for analysis; and
receiving, over the communication network and from the remote computer device, analysis results.

### Claim 2 (depends on 1)

The method of claim 1 wherein the analysis results include a set of numerical analysis-result value(s), the method further comprising:
for each given analysis-result value of the set of analysis result value(s), performing a transformation compensation on the given analysis-result value to yield a respectively corresponding anti-transformed analysis result value, with the transformation compensation being a set of mathematical operation(s) that cause the anti-transformed analysis result value to at least substantially match an analysis result value that would have been obtained if the first time-series data set had been sent for analysis.

### Claim 3 (depends on 2)

The method of claim 2 wherein the set of mathematical operation(s) of the transformation compensation are inverse mathematical operation(s) with respect to the set of mathematical operation(s) of the linear transformation.

### Claim 4 (depends on 1)

The method of claim 1 wherein:
the first time-series data is confidential business data; and
the analysis results include a business related prediction.

### Claim 5 (depends on 3)

The method of claim 3 further comprising:
taking a business related action based, at least in part, upon the business related prediction, with the business related action including one of the following types of business related actions: taking a staffing action, making an investment, foregoing an investment opportunity, making an inventory decision, and/or making a container allocation decision.

### Claim 6 (depends on 1)

The method of claim 1 wherein the set of mathematical operation(s) of the linear transformation includes addition of a constant value to all values of a time series.

### Claim 7 (depends on 1)

The method of claim 1 wherein the set of mathematical operation(s) of the linear transformation includes subtraction of a constant value from all values of a time series.

### Claim 8 (depends on 1)

The method of claim 1 wherein the set of mathematical operation(s) of the linear transformation includes multiplication by a constant value of all values of a time series.

### Claim 9 (depends on 1)

The method of claim 1 wherein the set of mathematical operation(s) of the linear transformation includes division by a constant value of all values of a time series.

### Claim 10 (depends on 1)

The method of claim 1 wherein the set of mathematical operation(s) of the linear transformation includes performing a first matrix algebra operation on the values of the first plurality of time-series values.

### Claim 11

A method comprising:
receiving a first time-series data set, with the first time-series data set including: (i) a plurality of time-series values, and (ii) a relative order of the first plurality of time-series values;
receiving a second time-series data set, with the second time-series data set including: (i) a plurality of time-series values, and (ii) a relative order of the second plurality of time-series values;
combining by a mathematical operation respectively corresponding pairs of time-series values from the first and second time-series data sets to obtain a combined time-series data set;
sending, over a communication network and to a remote computer device, the combined time-series data set for analysis; and
receiving, over the communication network and from the remote computer device, analysis results.

### Claim 12 (depends on 11)

The method of claim 11 wherein the analysis results include a set of numerical analysis-result value(s), the method further comprising:
for each given analysis-result value of the set of analysis result value(s), performing a combination compensation on the given analysis-result value to yield a respectively corresponding anti-combined analysis result value, with the combination compensation being a set of mathematical operation(s) that cause the anti-combined analysis result value to at least substantially match an analysis result values that would have been obtained if the first time-series data set and the second time-series data set had been separately sent for analysis.

### Claim 13 (depends on 11)

The method of claim 11 wherein:
the first time-series data is confidential business data;
the second time-series data is confidential business data; and
the analysis results include a business related prediction.

### Claim 14 (depends on 11)

The method of claim 11 further comprising:
taking a business related action based, at least in part, upon the business related prediction, with the business related action including one of the following types of business related actions: taking a staffing action, making an investment, foregoing an investment opportunity, making an inventory decision and/or making a container allocation decision.

### Claim 15 (depends on 11)

The method of claim 11 wherein the set of mathematical operation(s) of the linear transformation includes addition or subtraction.

### Claim 16 (depends on 11)

The method of claim 11 wherein the set of mathematical operation(s) of the linear transformation includes multiplication or division.

### Claim 17 (depends on 11)

The method of claim 11 further comprising:
receiving a third time-series data set, with the third time-series data set including: (i) a plurality of time-series values, and (ii) a relative order of the third plurality of time-series values;
wherein the combination by a mathematical operation further combines the time series values of the third time-series data sets with the time-series values from the first and second time-series data sets to obtain the combined time-series data set.

### Claim 18

A method comprising:
receiving a first time-series data set, with the first time-series data set including: (i) a plurality of time-series values, and (ii) a relative order of the first plurality of time-series values;
receiving a second time-series data set, with the second time-series data set including: (i) a plurality of time-series values, and (ii) a relative order of the second plurality of time-series values;
performing a first linear transformation on the first time-series data set to obtain a first transformed time series data set, with the linear transformation being a set of mathematical operation(s) that: (i) change each value of the first plurality of time-series values respectively corresponding transformed values, and (ii) maintains, in the transformed values, proportions that exist among and between the values of the first plurality of time-series values;
performing a second linear transformation on the first time-series data set to obtain a second transformed time series data set, with the linear transformation being a set of mathematical operation(s) that: (i) change each value of the first plurality of time-series values respectively corresponding transformed values, and (ii) maintains, in the transformed values, proportions that exist among and between the values of the first plurality of time-series values;
combining by a mathematical operation respectively corresponding pairs of time-series values from the first transformed and second transformed time-series data sets to obtain a combined time-series data set;
sending, over a communication network and to a remote computer device, the combined time-series data set for analysis; and
receiving, over the communication network and from the remote computer device, analysis results.

### Claim 19 (depends on 18)

The method of claim 18 wherein:
the set of mathematical operation(s) of the first linear transformation includes performing a first matrix algebra operation on the values of the first plurality of time-series values; and
the set of mathematical operation(s) of the second linear transformation includes performing a second matrix algebra operation on the values of the second plurality of time-series values.

### Claim 20 (depends on 18)

The method of claim 18 wherein the analysis results include a set of numerical analysis-result value(s), the method further comprising:
for each given analysis-result value of the set of analysis result value(s), performing a combination compensation on the given analysis-result value to yield a respectively corresponding anti-combined analysis result value, with the combination compensation being a set of mathematical operation(s) that cause the anti-combined analysis result value to at least substantially match an analysis result values that would have been obtained if the first transformed time-series data set and the second transformed time-series data set had been separately sent for analysis; and
for each given anti-combined result value of the set of anti-combined result value(s), performing a transformation compensation on the given analysis-result value to yield a respectively corresponding anti-transformed analysis result value, with the transformation compensation being a set of mathematical operation(s) that cause the anti-transformed analysis result value to at least substantially match an analysis result value that would have been obtained if the first time-series data set and second time-series data set had been sent for analysis.
