# US 2020/0402019 A1 · Techniques to apply machine learning to schedule events of interest

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

- Number: US 2020/0402019 A1
- Document ID: US20200402019A1
- Status: Published application
- Kind: A1
- Application number: 16/444,819
- Filed: 2019-06-18
- Published: 2020-12-24
- Earliest claimed: 2019-06-18
- Assignees: Capital One Services, LLC
- Applicants: Capital One Services, LLC
- Inventors: Anh TRUONG; Mark Louis WATSON; Austin Grant WALTERS; Jeremy Edward GOODSITT; Reza FARIVAR; Vincent PHAM; Fardin ABDI TAGHI ABAD; Kenneth TAYLOR
- Web page: https://patentgrep.com/patent/US20200402019A1

## CPC

- G06Q 10/1093: Calendar-based scheduling for persons or groups
- G06F 16/906: Clustering; Classification
- G06F 16/908: using metadata automatically derived from the content
- G06F 40/216: using statistical methods
- G06N 20/00: Machine learning
- 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
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Techniques to apply machine learning to schedule events of interest for a device user. As described herein, a typical device user is inundated with information suggesting activities and other things to do. Using these techniques, the information is categorized according to the activity and those activities that are most likely to be engaged in by the device user are recommended to the device user (via their device) as events of interest. If the device user selects an event of interest, the device is updated to reflect that selection. For instance, a calendar application may be updated to include an event description at the event's date and time. Other embodiments are described and claimed.

## Claims

### Claim 1

A device, comprising:
a processing circuit;
a computer memory storing information from a number of computer applications executed by the processing circuit; and
logic stored in the computer memory and executed on the processing circuit,
wherein the logic is operative to cause the processing circuit to:
identify in the number of applications executed by the processing circuit user activity data and user availability data from past behavior of a user of the device;
store the user activity data and the user availability data in the computer memory as database information;
access, from the computer memory, the database information including records corresponding to user activity data, user availability data, and events of interest in a geographic area, wherein the information is provided by the number of applications;
process the database information comprising the records corresponding to the events in the geographic area;
train a machine learning model having features based in part on user preference data in the records in the database information, the user preference data being based on user availability data and user activity data associated with various applications, the user activity data comprising historical events engaged in by a device user, each of the historic events comprising a historical activity attribute and a historical time attribute, the records comprising time attributes and activity attributes for events, and the features comprising a likelihood that a user will engage in an event with a time attribute corresponding to at least one of the historical time attributes of the historical events;
apply the trained machine learning model to the accessed database information related to the user activity data, the user availability data and the events of interest in the geographic area;
identify records having events of interest based upon values provided by the trained machine learning model by selecting the events of interest that include time attributes that are at future times, each of the events of interest comprising attributes that map to a value exceeding a threshold value and to corresponding attributes of at least a portion of the user preference data, wherein the identifying further includes selecting the attributes that map to the value exceeding the threshold value correspond to a likelihood determined by the machine learning model that the apparatus user will engage in the event of interest, and the likelihood related to the selected events of interest is based on determining at least one time attribute and at least one activity attribute of the selected events of interest exceeding the threshold value, and one or both of:
the event of interest having an activity attribute similar to a historical activity attribute of at least one of the historical events engaged in by the apparatus user, the at least one of the historical events having a historical time attribute corresponding with the time attribute of the event of interest; or
the event of interest having a time attribute similar to a historical time attribute of at least one of the historical events engaged in by the apparatus user, the at least one of the historical events having a historical activity attribute corresponding with the activity attribute of the event of interest;
generate recommendation data for one or more events of interest in the geographic area in the identified records;
display, on a display device of the apparatus, the recommendation data and a selectable option for each event of the one or more events; and
store one or more entries in a calendar application executed by the processing circuit based upon one or more user selections of one or more selectable options on the displayed recommendation data.

### Claim 2 (depends on 1)

The device of claim 1, comprising logic operative to cause the processing circuit to compare the identified records with the user availability data to determine whether there is an open time slot in the calendar application or a closed time slot in the calendar application for an event of the one or more events.

### Claim 3 (depends on 1)

The device of claim 1 comprising logic operative to cause the processing circuit to use the machine learning model to determine a likelihood that an event is an event of record and determine that an event is an event of interest based upon the likelihood.

### Claim 4 (depends on 1)

The device of claim 1 comprising logic operative to cause the processing circuit to update the machine learning model in response to the one or more user selections.

### Claim 5 (depends on 1)

The device of claim 1 comprising logic operative to cause the processing circuit to group into clusters of historical events of the user activity data.

### Claim 6 (depends on 1)

The device of claim 1 comprising logic operative to instruct a communication application to generate a communication directed towards an event leader.

### Claim 7 (depends on 1)

The device of claim 1 comprising logic operative to identify records having events of interest by comparing clusters of the user availability data with a date attribute, a time attribute, or both the date attribute and the time attribute.

### Claim 8 (depends on 1)

The device of claim 1 comprising logic operative to identify records having events of interest based upon a comparison between attribute data within a record and clusters of the user activity data being stored in the various applications.

### Claim 9

A computer-implemented method executed on a processing circuit in a mobile device, comprising:
identifying in the number of applications executed by the processing circuit user activity data and user availability data from past behavior of a user of the mobile device;
storing the user activity data and the user availability data in a memory of the mobile device as database information;
accessing, from the computer memory, database information including records corresponding to events in a geographic area, a mobile device user's activity data and the mobile device user's availability, wherein the information is provided by the number of applications;
processing the database information comprising records corresponding to the events in the geographic area;
training a machine learning model having features based in part on user preference data that correspond to clusters of historical events associated with the mobile device wherein each cluster corresponds to at least one attribute as defined in the records, the at least one attribute including a historical time attribute and a historical activity attribute, and the features including a probability that a user will engage in an event with a time attribute corresponding to the historical time attribute;
apply the trained machine learning model to the accessed database information related to the mobile device user's activity data, the device user's availability data and the events in the geographic area;
identifying records having events of interest of which each event of interest has a likelihood that exceeds a threshold value based upon the applying of the machine learning model by selecting the events of interest that include time attributes that are at future times, attributes that map to corresponding attributes of at least a portion of the user preference data, wherein the identifying further includes selecting at least one time attribute and at least one activity attribute of the attributes of the events of interest in the geographic area that map to a value exceeding the threshold value and that correspond to a likelihood determined by the machine learning model that the user will engage in the at least one event of interest in the geographic area, wherein the likelihood is further based on one or both of:
the event of interest having an activity attribute similar to a historical activity attribute of at least one of the clusters of historical events engaged in by the user, the at least one of the clusters of historical events having a historical time attribute corresponding with the time attribute of the event of interest; or
the event of interest having a time attribute similar to a historical time attribute of at least one of the clusters of historical events engaged in by the user, the at least one of the clusters of historical events having a historical activity attribute corresponding with the activity attribute of the event of interest;
generating recommendation data for an event in the identified records if there is an open time slot in a calendar application for that event;
displaying, on a display device, the recommendation data and a selectable option for the event in the recommendation data; and
storing an entry in the calendar application in response to a user selection of the selectable option for the event in the displayed recommendation data.

### Claim 10 (depends on 9)

The method of claim 9 further comprising updating the machine learning model in response to the user selection.

### Claim 11 (depends on 9)

The method of claim 9 further comprising comparing the identified records with user availability data to determine whether there is an open time slot in the calendar application or a closed time slot in the calendar application for an event of the one or more events.

### Claim 12 (depends on 9)

The method of claim 9 further comprising mapping portions of an event description to attributes corresponding to user availability data or user activity data.

### Claim 13 (depends on 9)

The method of claim 9 further comprising comparing the clusters of the historical events with at least one of a date attribute, a time attribute, a location attribute, an activity attribute, or an activity type attribute.

### Claim 14 (depends on 9)

The method of claim 9 further comprising determining that an event is an event of interest based upon a likelihood provided by a cluster corresponding to the at least one of a date attribute, a time attribute, a location attribute, an activity attribute, or an activity type attribute.

### Claim 15

At least one computer-readable storage medium comprising instructions that, when executed, cause a system to:
identify in the number of applications executed by the processing circuit user activity data and user availability data from past behavior of a user of the device;
store the user activity data and the user availability data in a computer memory as database information;
access, from the computer memory, the database information including records corresponding to user activity data, user availability data, and events of interest in a geographic area, wherein the information is provided by the number of applications; process database information comprising records corresponding to events in a geographic area;
train a machine learning model having features based in part on user preference data that comprises clusters of attribute data for historical events engaged in by a user, the user preference data being based on user availability data and user activity data associated with various applications, each cluster of the attribute data for historical events comprising a historical time attribute and a historical activity time attribute, and the features including a likelihood that the user will engage in an event having a time attribute corresponding to the historical time attribute;
identify a record having an event of interest based upon a likelihood associated with attribute data of the record, the event of interest comprising a time attribute that is at a future time, the likelihood being provided by the machine learning model, the likelihood based on one or both of:
the event of interest having an activity attribute similar to the historical activity attribute of a cluster of the clusters of attribute data, the cluster of the clusters of attribute data having a historical time attribute corresponding with the time attribute of the event of interest; or
the event of interest having a time attribute similar to the historical time attribute of a historical event of the cluster of the clusters of attribute data, the cluster of the clusters of attribute data having a historical activity attribute corresponding with the activity attribute of the event of interest,
wherein the identifying includes using the machine learning model to determine the likelihood associated with attribute data of the record based upon a cluster corresponding to attribute data of the event of interest;
compare the identified record with the user availability data to determine whether there is an open time slot in a calendar application or a closed time slot in the calendar application for the event in the identified record;
generate recommendation data for the event of interest based on a result of the comparison;
display, on a display device, the recommendation data and a selectable option for the event of interest; and
store one or more entries in the calendar application in response to a user selection of the selectable option on the displayed recommendation data.

### Claim 16

(canceled)

### Claim 17 (depends on 15)

The computer-readable storage medium of claim 15, comprising instructions that when executed cause the system to:
group into clusters of historical events of the user activity data.

### Claim 18

(canceled)

### Claim 19 (depends on 15)

The computer-readable storage medium of claim 15, comprising instructions that when executed cause the system to:
instruct a communication application to generate a communication directed towards an event leader.

### Claim 20 (depends on 15)

The computer-readable storage medium of claim 15, comprising instructions that when executed cause the system to:
store a contact in a contacts portions for an event leader.

### Claim 21 (depends on 1)

The device of claim 1, wherein identifying in the number of applications executed by the processing circuit user activity data and user availability data from past behavior of a user of the device, further comprises:
interpreting raw data and understanding content from the number of applications using a natural language processing mechanism or text summarization technique.

### Claim 22 (depends on 1)

The device of claim 1, wherein identifying in the number of applications executed by the processing circuit user activity data and user availability data from past behavior of a user of the device, further comprises:
analyzing textual information provided by an application of the number of applications; and
generating a vector of attributes including an activity attribute.
