Granted Patent B2

US 11,521,096 · App. 15/689,757 · Granted Dec 6, 2022

Inventors: Daniel Erenrich (Mountain View, CA); Anirvan Mukherjee (Mountain View, CA)

Assignee: Palantir Technologies Inc.

G06N5/048G06F16/2455G06N7/005G06Q30/01G06Q30/0202G06Q30/0201

Abstract

Claims (50)

1. A system comprising:

one or more processors; and

memory storing instructions that, when executed by the one or more processors, cause the system to:

obtain information associated with a plurality of entities from one or more data sources;

form a plurality of records associated with the plurality of entities by integrating the information from the one or more data sources, each of the plurality of records associated with a corresponding entity of the plurality of entities and comprising a plurality of data entries from a specific data source of the one or more data sources, wherein each of the plurality of records stores the plurality of data entries serially;

filter the plurality of records for a subset of the information associated with a specified action and a specified temporal period;

generate, based on the filtering, a record label for each of the plurality of records;

generate a key value pair for each of the plurality of records, wherein the key value pair comprises a feature and a corresponding value, the feature indicating the corresponding value is an average value, and the corresponding value comprising an average value of the plurality of serially stored data entries from the specific data source;

create a set of training data items, each training data item of the set of training data items comprising a corresponding key value pair of the generated key value pairs and a corresponding record label of the generated record labels;

select a subset of the training data items;

obtain a model using the subset of training data items, wherein the obtaining comprises constructing multiple models using the subset of training data items, each of the multiple models having a plurality of performance metrics, and selecting one of the multiple models based on one or more of the plurality of performance metrics;

determine, using the model, a propensity of a particular entity of the plurality of entities to perform a specified action; and

generate a graphical user interface indicating the propensity of the particular entity of the plurality of entities to perform a specified action.

2. The system of claim 1 , wherein each of the record labels indicate whether the corresponding entity performed the specified action within the specified temporal period.

3. The system of claim 1 , wherein the instructions further cause the system to determine, based on the trained model and a corresponding record of the plurality of records, a relative importance of the features.

4. The system of claim 1 , wherein the graphical user interface comprises a plurality of tiles, each of the plurality of tiles being associated with a corresponding entity of a subset of the plurality of entities.

5. The system of claim 4 , wherein a particular tile of the plurality of tiles is associated with the particular entity, and a placement of the particular tile of the plurality of tiles within the graphical user interface is based on the determined propensity of the particular entity to perform the specified action.

6. The system of claim 5 , wherein the instructions further cause the system to update the placement of the particular tile of the plurality of tiles within the graphical user interfaces in response to an update of the information.

7. A method for determining a propensity of an entity to take a specified action, the method being performed by one or more processors and comprising:

obtaining information associated with a plurality of entities from one or more data sources;

forming a plurality of records associated with the plurality of entities by integrating the information from the one or more data sources, each of the plurality of records associated with a corresponding entity of the plurality of entities and comprising a plurality of data entries from a specific data source of the one or more data sources, wherein each of the plurality of records stores the plurality of data entries serially;

filtering the plurality of records for a subset of the information associated with a specified action and a specified temporal period;

generating, based on the filtering, a record label for each of the plurality of records;

generating a key value pair for each of the plurality of records, wherein the key value pair comprises a feature and a corresponding value, the feature indicating the corresponding value is an average value, and the corresponding value comprising an average value of the plurality of serially stored data entries from the specific data source;

creating a set of training data items, each training data item of the set of training data items comprising a corresponding key value pair of the generated key value pairs and a corresponding record label of the generated record labels;

selecting a subset of the training data items;

obtaining a model using the subset of training data items, wherein the obtaining comprises constructing multiple models using the subset of training data items, each of the multiple models having a plurality of performance metrics, and selecting one of the multiple models based on one or more of the plurality of performance metrics;

determining, using the model, a propensity of a particular entity of the plurality of entities to perform a specified action; and

generating a graphical user interface indicating the propensity of the particular entity of the plurality of entities to perform a specified action.

8. The method of claim 7 , wherein each of the record labels indicate whether the corresponding entity performed the specified action within the specified temporal period.

9. The method of claim 7 , further comprising determining, based on the trained model and a corresponding record of the plurality of records, a relative importance of the features.

10. The method of claim 7 , wherein the graphical user interface comprises a plurality of tiles, each of the plurality of tiles being associated with a corresponding entity of a subset of the plurality of entities.

11. The method of claim 10 , wherein a particular tile of the plurality of tiles is associated with the particular entity, and a placement of the particular tile of the plurality of tiles within the graphical user interface is based on the determined propensity of the particular entity to perform the specified action.

12. The method of claim 11 , further comprising updating the placement of the particular tile of the plurality of tiles within the graphical user interfaces in response to an update of the information.

13. A non-transitory computer-readable medium storing a set of instructions that are executable by one or more processors to cause the one or more processors to perform a method, the method comprising:

obtaining information associated with a plurality of entities from one or more data sources;

forming a plurality of records associated with the plurality of entities by integrating the information from the one or more data sources, each of the plurality of records associated with a corresponding entity of the plurality of entities and comprising a plurality of data entries from a specific data source of the one or more data sources, wherein each of the plurality of records stores the plurality of data entries serially;

filtering the plurality of records for a subset of the information associated with a specified action and a specified temporal period;

generating, based on the filtering, a record label for each of the plurality of records;

generating a key value pair for each of the plurality of records, wherein the key value pair comprises a feature and a corresponding value, the feature indicating the corresponding value is an average value, and the corresponding value comprising an average value of the plurality of serially stored data entries from the specific data source;

creating a set of training data items, each training data item of the set of training data items comprising a corresponding key value pair of the generated key value pairs and a corresponding record label of the generated record labels;

selecting a subset of the training data items;

obtaining a model using the subset of training data items, wherein the obtaining comprises constructing multiple models using the subset of training data items, each of the multiple models having a plurality of performance metrics, and selecting one of the multiple models based on one or more of the plurality of performance metrics;

determining, using the model, a propensity of a particular entity of the plurality of entities to perform a specified action; and

generating a graphical user interface indicating the propensity of the particular entity of the plurality of entities to perform a specified action.

14. The non-transitory computer-readable medium of claim 13 , wherein each of the record labels indicate whether the corresponding entity performed the specified action within the specified temporal period.

15. The non-transitory computer-readable medium of claim 13 , wherein the graphical user interface comprises a plurality of tiles, each of the plurality of tiles being associated with a corresponding entity of a subset of the plurality of entities.

16. The non-transitory computer-readable medium of claim 15 , wherein a particular tile of the plurality of tiles is associated with the particular entity, and a placement of the particular tile of the plurality of tiles within the graphical user interface is based on the determined propensity of the particular entity to perform the specified action.

17. The non-transitory computer-readable medium of claim 16 , further comprising instructions executable by the one or more processors to cause the one or more processors to perform:

updating the placement of the particular tile of the plurality of tiles within the graphical user interfaces in response to an update of the information.

Assignments (8)

Continuity (4)