# US 2019/0318125 A1 · Augmenting datasets with selected de-identified data records

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

- Number: US 2019/0318125 A1
- Document ID: US20190318125A1
- Status: Published application
- Kind: A1
- Application number: 16/449,682
- Filed: 2019-06-24
- Published: 2019-10-17
- Earliest claimed: 2018-04-12
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Aris Gkoulalas-Divanis
- Web page: https://patentgrep.com/patent/US20190318125A1

## CPC

- G06F 21/6254: {by anonymising data, e.g. decorrelating personal data from the owner's identification}
- G06F 21/604: {Tools and structures for managing or administering access control systems}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A computer system utilizes a dataset to support a research study. Regions of interestingness are determined within a model of data records of a first dataset that are authorized for the research study by associated entities. Data records from a second dataset are represented within the model, wherein the data records from the second dataset are relevant for supporting objectives of the research study. Data records from the second dataset that fail to satisfy de-identification requirements are removed. A resulting dataset is generated that including the first dataset records within a selected region of interestingness and selected records of the second dataset within the same region. The second dataset records within the resulting dataset are de-identified based on the de-identification requirements. Embodiments of the present invention further include a method and program product for utilizing a dataset to support a research study in substantially the same manner described above.

## Claims

### Claim 1

A method, in a data processing system comprising at least one processor and at least one memory, the at least one memory comprising instructions executed by the at least one processor to cause the at least one processor to utilize a dataset for supporting a research study, the method comprising:
determining one or more regions of interestingness within a model of data records of a first dataset, wherein the data records of the first dataset are authorized for the research study by associated entities;
representing within the model data records from a second dataset, wherein the data records from the second dataset are relevant for supporting objectives of the research study, correspond to entities other than those associated with the first dataset, and are authorized for the research study by associated entities after transformation to satisfy de-identification requirements;
removing data records of the second dataset from the one or more regions of interestingness based on those records failing to satisfy the de-identification requirements;
generating a resulting dataset for the research study including the data records of the first dataset within a selected region of interestingness and selected data records of the second dataset within the selected region of interestingness; and
de-identifying the data records of the second dataset within the resulting dataset based on the de-identification requirements.

### Claim 2 (depends on 1)

The method of claim 1, wherein generating a resulting dataset further comprises:
identifying data records of the second dataset within the selected region of interestingness that are similar to data records of the first dataset; and
generating the resulting dataset with the identified data records of the second dataset.

### Claim 3 (depends on 1)

The method of claim 1, wherein the data records from the second data set that are relevant for supporting objectives of the research comprise data records from the second data set that fit within a region of interestingness.

### Claim 4 (depends on 1)

The method of claim 1, wherein the data records of the second dataset are de-identified in separate groups corresponding to each of the one or more regions of interestingness.

### Claim 5 (depends on 1)

The method of claim 1, wherein each region of interestingness is identified by analyzing the data records from the first dataset to find clusters of records.