Patent Issued for Systems for predicting and classifying location data based on machine learning (USPTO 11297466): Allstate Insurance Company
2022 APR 21 (NewsRx) -- By a
The patent’s inventors are Nies, Anne (
This patent was filed on
From the background information supplied by the inventors, news correspondents obtained the following quote: “A person may change his or her residence, for instance by moving from one geographic location to another. It may be desirable to proactively identify a move (e.g., without user input) in order to execute one or more modifications to user data to maintain accuracy.”
Supplementing the background information on this patent, NewsRx reporters also obtained the inventors’ summary information for this patent: “In light of the foregoing background, the following presents a simplified summary of the present disclosure in order to provide a basic understanding of some aspects of the invention. This summary is not an extensive overview of the invention. It is not intended to identify key or critical elements of the invention or to delineate the scope of the invention. The following summary merely presents some concepts of the invention in a simplified form as a prelude to the more detailed description provided below.
“Currently, it is difficult to predict if or when a person will change his or her residence. As a result, it is not currently possible to predictively identify a move and/or update user data records. Accordingly, there is a need to accurately predict if and when a person is planning to move his or her residence. Systems and methods described herein can use a variety of computing devices to obtain location data that can be used to generate a prediction (e.g., based on machine learning) of a likelihood that a person will move his or her residence. The location data can be generated based on location data captured from a mobile computing device and/or based on telematics data captured during the operation of a vehicle. The location measurement may include global navigation satellite system (GNSS) data, such as Global Positioning System (GPS) data generated by a GPS receiver or may comprise other location data, such as mapping information captured by one or more applications executing on a computing device. A computing device may compile the location data into graphs comprising edges and nodes. The nodes may represent locations visited by the person. The edges may represent a relationship between two of the locations visited by the person, such as instances of the person traveling between the two locations.
“In several embodiments, machine classifiers are trained based on training data comprising location data and/or graph data representing instances of ground truth in which a person changed the location of his or her residence. After a machine classifier has been trained, the machine classifier may take a graph representation of location data as input data, and, based on the input data, may output a prediction of a likelihood that a person will change his or her residence.
“The arrangements described can also include other additional elements, steps, computer-executable instructions, or computer-readable data structures. In this regard, other embodiments are disclosed and claimed herein as well. The details of these and other embodiments of the present invention are set forth in the accompanying drawings and the description below. Other features and advantages of the invention will be apparent from the description, drawings, and claims.”
The claims supplied by the inventors are:
“1. A method for determining a change in residence, comprising: obtaining, by a computing device, location data indicating a set of geographic locations and a set of times that a person was at each of the geographic locations; generating, by the computing device and based on the location data, a first graph data comprising a first set of nodes and edges therebetween, and a second graph data comprising a second set of nodes and edges therebetween; ranking, by the computing device and based on amounts of time the person spent at a first set of geographic locations associated with the first graph data, the first set of nodes to produce a first set of ranked nodes; ranking, by the computing device and based on amounts of time the person spent at a second set of geographic locations associated with the second graph, the second set of nodes to produce a second set of ranked nodes; determining, by the computing device, geographic distances between the first set of geographic locations corresponding to the first set of ranked nodes and the second set of geographic locations corresponding to the second set of ranked nodes; and determining, by the computing device and based on the geographic distances, a predicted likelihood that the person will move a residence, wherein the predicted likelihood is determined using a machine learning classifier.
“2. The method of claim 1, wherein determining the predicted likelihood comprises: summing, by the computing device, the geographic distances to produce a sum of the geographic distances; and determining, by the computing device, the predicted likelihood based on the sum of the geographic distances.
“3. The method of claim 1, further comprising: obtaining, by the computing device, a ground truth label indicating if the person moved the residence; and retraining, by the computing device, the machine learning classifier based on the geographic distance, the predicted likelihood, and the ground truth label.
“4. The method of claim 1, wherein the machine classifier comprises a logistic classifier.
“5. The method of claim 1, further comprising: regularizing, by the computing device, the geographic distances between the geographic locations to produce regularized distances; and determining the predicted likelihood that the person will move the residence further comprises classifying a sum of the regularized distances to determine the predicted likelihood.
“6. The method of claim 1, further comprising determining a frequency that the person visited the first set of geographic locations and a frequency that the person visited the second set of geographic locations, wherein the predicted likelihood that the person will move the residence is based on the frequency that the person visited the first set of geographic locations and the frequency that the person visited the second set of geographic locations.
“7. The method of claim 1, wherein determining the predicted likelihood that the person will move the residence is based on the first set of ranked nodes and the second set of ranked nodes.
“8. The method of claim 1, wherein the geographic distances comprise haversine distances.
“9. The method of claim 1, further comprising: obtaining data, by the computing device and from a data source, related to a likelihood that the person will move the residence; and determining the predicted likelihood further comprises determining the predicted likelihood based on the geographic distances and the obtained data.
“10. The method of claim 1, wherein the first graph data comprises the first set of nodes and edges, each of the first set of nodes corresponds to a geographic location of the first set of the geographic locations that the person visited during a first time period, and each of the first set of edges indicates a connection between two of the first set of the geographic locations; and wherein the second graph comprises the second set of nodes and edges, each of the second set of nodes corresponds to a geographic location of the second set of the geographic locations that the person visited during a second time period, and each of the second set of edges indicates a connection between two of the second set of locations.
“11. A computing device for determining a change in residence, comprising: a processor; and memory storing instructions that, when executed by the processor, cause the computing device to: obtain location data indicating a set of geographic locations and a set of times that a person was at each of the geographic locations; generate a first graph data comprising a first set of nodes and edges therebetween, and a second graph data comprising a second set of nodes and edges therebetween; rank, based on amounts of time the person spent at a first set of geographic locations associated with the first graph data, the first set of nodes to produce a first set of ranked nodes; rank, based on amounts of time the person spent at a second set of geographic locations associated with the second graph, the second set of nodes to produce a second set of ranked nodes; determine geographic distances between the first set of geographic locations corresponding to the first set of ranked nodes and the second set of geographic locations corresponding to the second set of ranked nodes; determine, based on the geographic distances and with a machine classifier, a predicted likelihood that the person will move a residence; obtain a ground truth label indicating if the person moved the residence; and retrain the machine classifier based on the geographic distance, the determined predicted likelihood, and the ground truth label.
“12. The computing device of claim 11, wherein the instructions that cause the computing device to determine the predicted likelihood further comprise instructions that, when executed by the one or more processors, further cause the computing device to: sum the geographic distances to produce a sum of the geographic distances; and determine the predicted likelihood based on the sum of the geographic distances.
“13. The computing device of claim 11, wherein the machine classifier comprises a logistic classifier.
“14. The computing device of claim 11, wherein the instructions, when executed by the one or more processors, further cause the computing device to: regularize the geographic distances between the first set of geographic locations corresponding to the first set of ranked nodes and the second set of geographic locations corresponding to the second set of ranked nodes to produce a set of regularized distances, and classify a sum of the regularized distances to determine the predicted likelihood.
“15. The computing device of claim 11, wherein the instructions, when executed by the one or more processors, further cause the computing device to determine a frequency that the person visited the first set of geographic locations and a frequency that the person visited the second set of geographic locations, wherein the predicted likelihood that the person will move the residence is based on the frequency that the person visited the first set of geographic locations and the frequency that the person visited the second set of geographic locations.
“16. The computing device of claim 11, wherein the instructions, when executed by the one or more processors, further cause the computing device to determine the predicted likelihood that the person will move the residence based on the first set of ranked nodes and the second set of ranked nodes.
“17. A non-transitory machine-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform steps comprising: obtaining location data indicating a set of geographic locations and a set of times that a person was at each of the geographic locations; generating a first graph data comprising a first set of nodes and edges therebetween, and a second graph data comprising a second set of nodes and edges therebetween; ranking, based on amounts of time the person spent at a first set of geographic locations associated with the first graph data, the first set of nodes to produce a first set of ranked nodes; ranking, based on amounts of time the person spent at a second set of geographic locations associated with the second graph, the second set of nodes to produce a second set of ranked nodes; determining haversine distances between the first set of geographic locations corresponding to the first set of ranked nodes and the second set of geographic locations corresponding to the second set of ranked nodes; determining, based on the haversine distances and with a logistic machine classifier, a predicted likelihood that the person will move a residence; obtaining a ground truth label indicating if the person moved the residence; and retraining the machine classifier based on the haversine distance, the predicted likelihood, and the ground truth label.
“18. The non-transitory machine-readable medium of claim 17, wherein the instructions, when executed by one or more processors, further cause the one or more processors to determine the predicted likelihood by: summing the haversine distances to produce a sum of the haversine distances; and determining the predicted likelihood based on the sum of the haversine distances.
“19. The non-transitory machine-readable medium of claim 17, wherein the predicted likelihood that the person will move the residence is based on the first set of ranked nodes and the second set of ranked nodes.
“20. The non-transitory machine-readable medium of claim 17, wherein the instructions, when executed by the one or more processors, cause the one or more processors to determine the predicted likelihood based on a frequency of time the person spent at the first set of geographic locations and a frequency of time the person spent at the second set of geographic locations.”
For the URL and additional information on this patent, see: Nies, Anne. Systems for predicting and classifying location data based on machine learning.
(Our reports deliver fact-based news of research and discoveries from around the world.)
Findings from University of Oslo Has Provided New Data on Managed Care (Local Public Health Projectification In Practice: a Qualitative Study of Facilitators and Barriers To a Public Health Plan Implementation): Managed Care
Patent Issued for Reward system related to a vehicle-to-vehicle communication system (USPTO 11295391): Allstate Insurance Company
Advisor News
- Year-end planning errors retirees should avoid
- Social cultivation: How to talk with wealthy prospects
- How important is a written financial plan?
- Coalition wants to advise people impacted by policy changes in new administration
- Diagnosing miscommunication in retirement planning
More Advisor NewsAnnuity News
Health/Employee Benefits News
- Interview | Acting Healthcare Advocate Kathy Holt
- Trump's CDC nominee is the reason pro-life values can't cost health care providers funding
- Health Insurance Sign-Up
- CVS Health cuts 42 more jobs linked to Aetna HQ in Hartford
- Molina Healthcare of Texas Hosts "Parents as Caregivers" Event in Weslaco
More Health/Employee Benefits NewsLife Insurance News