Systems and methods for implementing a machine learning approach to modeling entity behavior
Systems and methods are provided for implementing a machine learning approach to modeling entity behavior. Fixed information and periodically updated information may be utilized to predict the behavior of an entity. By incorporating periodically updated information, the system is able to maintain an up-to-date prediction of each entity's behavior, while also accounting for entity action with respect to ongoing obligations. The system may generate behavior scores for the set of entities. In some embodiments, the behavior scores that are generated may indicate the transactional risk associated with each entity. Using the behavior scores generated, a user may be able to assess the credit riskiness of individual entities and instruct one or more individuals assigned to the entities to take one or more actions based on the credit riskiness of the individual entities.
1. A system for implementing a machine learning approach to modeling entity behavior, the system comprising:
one or more processors; and
a memory storing instructions that, when executed by the one or more processors, cause the system to:
obtain first data for a set of entities related to a first time period;
label a subset of the first data based on identified classifications for one or more of the set of entities;
assign numerical labels based on the identified classifications;
obtain second data for the set of entities related to a second time period;
obtain third data for the set of entities related to a third time period;
identify dependencies between the second data and the third data;
update the second data based on the identified dependencies between the second data and the third data;
apply the machine learning model to the updated second data and the third data, wherein the machine learning model is trained to output behavior scores based on the numerical labels; and
determine the behavior scores for each of the set of entities based on the application of the machine learning model to the updated second data and the third data.
2. The system of claim 1 , wherein the second data comprises a plurality of entries, wherein the system is further caused to:
identify a set of correlations between each of the plurality of entries and one or more data types; and
generate a dataset comprising the plurality of entries based on the set of correlations, wherein the machine learning model is applied to the generated dataset.
3. The system of claim 2 , wherein each of the plurality of entries includes one or more features, wherein to generate the dataset comprising the plurality of entries, the system is further caused to:
identify a single data type associated with each feature;
obtain an indication of a predefined format for each data type; and
convert each feature to the predefined format for the data type to which it is associated.
4. The system of claim 1 , wherein the system is further caused to:
obtain access to multiple databases, wherein the multiple databases include a first database comprising fixed information for the set of entities and a second database comprising periodically updated information associated with the set of entities; and
obtain the first data from the multiple databases, wherein the first data comprises fixed information for the set of entities and periodically updated information for the set of entities related to the first time period.
5. The system of claim 1 , wherein the system is further caused to:
compare the second data and the third data;
identify one or more discrepancies between the second data and the third data based on the comparison;
identify changes that have occurred involving one or more of the set of entities based on the identified one or more discrepancies; and
generate an other dataset based on the identified changes.
6. The system of claim 1 , wherein to identify the dependencies between the second data and the third data, the system is further caused to:
obtain a dependency graph indicating one or more dependencies between data related to a second time period and data related to a third time period; and
identify the dependencies between the second data and the third data based on the dependency graph.
7. The system of claim 1 , wherein to identify the dependencies between the second data and the third data, the system is further caused to:
obtain a dependency graph indicating one or more dependencies between data from disparate datasets; and
identify the dependencies between the second data and the third data based on the dependency graph.
8. The system of claim 1 , wherein the machine learning model comprises a random forest model.
9. The system of claim 1 , wherein the training the machine learning model comprises determining whether to apply Gini impurity, information gain, or mean-squared-error as a splitting criteria to split a decision tree.
10. The system of claim 1 , wherein the machine learning model is based on modelling the set of entities as spokes linked to an anchor.
11. A method being implemented by a computing system having one or more processors and storage media storing machine-readable instructions that, when executed by the one or more processors, cause the computer system to perform the method, the method comprising:
obtaining first data for a set of entities related to a first time period;
labeling a subset of the first data based on identified classifications for one or more of the set of entities;
assigning numerical labels based on the identified classifications;
obtaining second data for the set of entities related to a second time period;
obtaining third data for the set of entities related to a third time period;
identifying dependencies between the second data and the third data;
updating the second data based on the identified dependencies between the second data and the third data;
applying the machine learning model to the updated second data and the third data, wherein the machine learning model is trained to output behavior scores based on the numerical labels; and
determining the behavior scores for each of the set of entities based on the application of the machine learning model to the updated second data and the third data.
12. The method of claim 11 , wherein the second data comprises a plurality of entries, the method further comprising:
identifying a set of correlations between each of the plurality of entries and one or more data types; and
generating a dataset comprising the plurality of entries based on the set of correlations, wherein the machine learning model is applied to the generated dataset.
13. The method of claim 12 , wherein each of the plurality of entries includes one or more features, wherein generating the dataset comprising the plurality of entries comprises:
identifying a single data type associated with each feature;
obtaining an indication of a predefined format for each data type; and
converting each feature to the predefined format for the data type to which it is associated.
14. The method of claim 11 , the method further comprising:
obtaining access to multiple databases, wherein the multiple databases include a first database comprising fixed information for the set of entities and a second database comprising periodically updated information associated with the set of entities; and
obtaining the first data from the multiple databases, wherein the first data comprises fixed information for the set of entities and periodically updated information for the set of entities related to the first time period.
15. The method of claim 11 , the method further comprising:
comparing the second data and the third data;
identifying one or more discrepancies between the second data and the third data based on the comparison;
identifying changes that have occurred involving one or more of the set of entities based on the identified one or more discrepancies; and
generating an other dataset based on the identified changes.
16. The method of claim 11 , wherein identifying the dependencies between the second data and the third data comprises:
obtaining a dependency graph indicating one or more dependencies between data related to a second time period and data related to a third time period; and
identifying the dependencies between the second data and the third data based on the dependency graph.
17. The method of claim 11 , wherein identifying the dependencies between the second data and the third data comprises:
obtaining a dependency graph indicating one or more dependencies between data from disparate datasets; and
identifying the dependencies between the second data and the third data based on the dependency graph.
18. The method of claim 11 , wherein the machine learning model comprises a random forest model.
19. The method of claim 11 , wherein the training the machine learning model comprises determining whether to apply Gini impurity, information gain, or mean-squared-error as a splitting criteria to split a decision tree.
20. A non-transitory computer readable medium comprising instructions that, when executed, cause one or more processors to perform:
obtaining first data for a set of entities related to a first time period;
labeling a subset of the first data based on identified classifications for one or more of the set of entities;
assigning numerical labels based on the identified classifications;
obtaining second data for the set of entities related to a second time period;
obtaining third data for the set of entities related to a third time period;
identifying dependencies between the second data and the third data;
updating the second data based on the identified dependencies between the second data and the third data;
applying the machine learning model to the updated second data and the third data, wherein the machine learning model is trained to output behavior scores based on the numerical labels; and
determining the behavior scores for each of the set of entities based on the application of the machine learning model to the updated second data and the third data.