Provider performance scoring using supervised and unsupervised learning
A system and a method are disclosed for a tool that generates a provider score corresponding to a predicted performance of a provider based on data of claims involving the provider. For a given claim, the tool provides the data as input into a supervised machine learning model and receives as output from the supervised machine learning model a predicted performance of the claim. The tool also inputs the data of the claim into an unsupervised machine learning model that is selected based on a stage of claim processing that the claim belongs to and receives as output from the unsupervised machine learning model an identification of a cluster of candidate claims to which the claim belongs. The tool combines the outputs of the supervised machine learning model and the unsupervised machine learning model to generate the provider score.
1 . A method for predicting performance of a provider comprising:
receiving data associated with a claim involving the provider;
inputting the data associated with the claim into a supervised machine learning model and receiving a predicted performance of the claim as output, wherein the supervised machine learning model has been trained on structured data and unstructured data;
selecting a first unsupervised machine learning model from a plurality of unsupervised machine learning models by determining whether information available for the claim satisfies data fields associated with a stage in claim processing that the claim belongs to, one or more of the plurality of unsupervised machine learning models associated with a different stage of a plurality of stages in claim processing, wherein the plurality of stages comprise information associated with the claim as the claim progresses;
inputting the data associated with the claim into the selected first unsupervised machine learning model and receiving as output from the selected first unsupervised machine learning model an identification of a cluster of candidate claims to which the claim belongs;
generating a score corresponding to a predicted performance of the provider based on the predicted performance of the claim and the identification of the cluster of candidate claims; and
providing, for display at a client device, the generated score of the provider.
2 . The method of claim 1 , further comprising:
receiving data associated with the provider; and
inputting the data associated with the provider and the data associated with the claim into a second unsupervised machine learning model from the plurality of unsupervised machine learning models and receiving as output from the second unsupervised machine learning model an identification of a cluster of candidate providers to which the provider belongs.
3 . The method of claim 2 , wherein generating the score corresponding to the predicted performance of the provider is further based on the identification of the cluster of candidate providers.
4 . The method of claim 1 , further comprising:
determining a contribution of the provider to the claim relative to other providers associated with the claim.
5 . The method of claim 1 , wherein the supervised machine learning model is further trained using a set of training data including data associated with historical claims, wherein one or more of the historical claims are associated with an enterprise to which the client device belongs.
6 . The method of claim 1 , wherein inputting the data associated with the claim into the selected first unsupervised machine learning model further comprises:
identifying a subset of the data associated with the stage in claim processing that the claim belongs to; and
inputting the subset of the data into the selected first unsupervised machine learning model.
7 . The method of claim 1 , wherein the generated score of the provider is presented relative to candidate providers in a cluster of candidate providers to which the provider belongs.
8 . The method of claim 1 , further comprising:
receiving additional data associated with the claim involving the provider, the additional data associated with a subsequent stage in claim processing;
responsive to receiving the additional data, selecting a second unsupervised machine learning model from the plurality of unsupervised machine learning models based on the subsequent stage that the claim belongs to;
inputting the additional data associated with the claim into the second unsupervised machine learning model and receiving as output from the second unsupervised machine learning model an identification of another cluster of candidate claims to which the claim belongs;
updating the score corresponding to a predicted performance of the provider based on the predicted performance of the claim and the identification of the other cluster of candidate claims; and
providing, for display at a client device, the updated score of the provider.
9 . The method of claim 1 , wherein training the supervised machine learning model comprises parsing the structured data and unstructured data.
10 . The method of claim 1 , wherein training the supervised machine learning model comprises reconciling a dimensionality of the structured and a dimensionality of the unstructured data.
11 . The method of claim 1 , wherein selecting the first unsupervised machine learning model from a plurality of unsupervised machine learning models reduces a latent space associated with the cluster of candidate claims to which the claim belongs.
12 . The method of claim 1 , wherein the information available for the claim comprises information from initial claim forms, indemnity related information associated, or bill line data, or combinations thereof.
13 . A non-transitory computer-readable medium comprising computer program instructions that, when executed by a computer processor, cause the processor to perform operations, the instructions comprising instructions to:
receive data associated with a claim involving a provider;
input the data associated with the claim into a supervised machine learning model and receiving predicted performance of the claim as output, wherein the supervised machine learning model is further trained using a set of training data including data associated with historical claims, wherein one or more of the historical claims are associated with an enterprise to which a client device belongs;
select a first unsupervised machine learning model from a plurality of unsupervised machine learning models by determining whether information available for the claim satisfies data fields associated with a stage in claim processing that the claim belongs to, one or more of the plurality of unsupervised machine learning models associated with a different stage of a plurality of stages in claim processing;
input the data associated with the claim into the selected first unsupervised machine learning model and receiving as output from the selected first unsupervised machine learning model an identification of a cluster of candidate claims to which the claim belongs;
generate a score corresponding to a predicted performance of the provider based on the predicted performance of the claim and the identification of the cluster of candidate claims; and provide, for display at a client device, the generated score of the provider.
14 . The non-transitory computer-readable medium of claim 13 , wherein the instructions comprise instructions to:
receive data associated with the provider; and
input the data associated with the provider and the data associated with the claim into a second unsupervised machine learning model from the plurality of unsupervised machine learning models and receiving as output from the second unsupervised machine learning model an identification of a cluster of candidate providers to which the provider belongs.
15 . The non-transitory computer-readable medium of claim 14 , wherein the instructions to generate the score corresponding to the predicted performance of the provider is further based on the identification of the cluster of candidate providers.
16 . The non-transitory computer-readable medium of claim 13 , wherein instructions comprise instructions to:
determine a contribution of the provider to the claim relative to other providers associated with the claim.
17 . The non-transitory computer-readable medium of claim 13 , wherein the instructions to input the data associated with the claim into the selected first unsupervised machine learning model further comprises instructions to:
identify a subset of the data associated with the stage in claim processing that the claim belongs to; and
input the subset of the data into the selected first unsupervised machine learning model.
18 . The non-transitory computer-readable medium of claim 13 , wherein the generated score of the provider is presented relative to candidate providers in a cluster of candidate providers to which the provider belongs.
19 . The non-transitory computer-readable medium of claim 13 , wherein the instructions comprise instructions to:
receive additional data associated with the claim involving the provider, the additional data associated with a subsequent stage in claim processing;
responsive to receiving the additional data, select a second unsupervised machine learning model from the plurality of unsupervised machine learning models based on the subsequent stage that the claim belongs to;
input the additional data associated with the claim into the second unsupervised machine learning model and receiving as output from the second unsupervised machine learning model an identification of a second cluster of candidate claims to which the claim belongs;
update the score corresponding to a predicted performance of the provider based on the predicted performance of the claim and the identification of the second cluster of candidate claims; and
providing, for display at a client device, the updated score of the provider.
20 . A system comprising:
a processor; and
a non-transitory computer-readable medium comprising computer program instructions that, when executed by a computer processor, cause the processor to perform operations comprising:
training a supervised machine learning model on structured data and unstructured data;
receiving data associated with a claim involving a provider;
inputting the data associated with the claim into the supervised machine learning model and receiving predicted performance of the claim as output;
selecting a first unsupervised machine learning model from a plurality of unsupervised machine learning models by determining whether information available for the claim satisfies data fields associated with a stage in claim processing that the claim belongs to, one or more of the plurality of unsupervised machine learning models associated with a different stage of a plurality of stages in claim processing, wherein the plurality of stages comprise information submitted by providers associated with the claim as the claim progresses;
inputting the data associated with the claim into the selected unsupervised machine learning model and receiving as output from the selected first unsupervised machine learning model an identification of a cluster of candidate claims to which the claim belongs;
generating a score corresponding to a predicted performance of the provider by combining the predicted performance of the claim and the identification of the cluster of candidate claims; and
providing, for display at a client device, the generated score of the provider.
21 . The system of claim 20 , wherein the computer program instructions cause the processor to further perform operations comprising:
receiving data associated with the provider; and
inputting the data associated with the provider and the data associated with the claim into a second unsupervised machine learning model from the plurality of unsupervised machine learning models and receiving as output from the second unsupervised machine learning model an identification of a cluster of candidate providers to which the provider belongs.
22 . The system of claim 20 , wherein the computer program instructions cause the processor to further perform operations comprising:
determining a contribution of the provider to the claim relative to other providers associated with the claim.
23 . The system of claim 20 , wherein the computer program instructions cause the processor to further perform operations comprising:
receiving additional data associated with the claim involving the provider, the additional data associated with a subsequent stage in claim processing;
responsive to receiving the additional data, selecting a second unsupervised machine learning model from the plurality of unsupervised machine learning models based on the subsequent stage that the claim belongs to;
inputting the additional data associated with the claim into the second unsupervised machine learning model and receiving as output from the second unsupervised machine learning model an identification of second cluster of candidate claims to which the claim belongs;
updating the score corresponding to a predicted performance of the provider based on the predicted performance of the claim and the identification of the second cluster of candidate claims; and
providing, for display at a client device, the updated score of the provider.
24 . A method for predicting performance of a provider comprising:
receiving data associated with a claim involving the provider;
inputting the data associated with the claim into a supervised machine learning model and receiving a predicted performance of the claim as output, wherein the supervised machine learning model has been trained on structured data and unstructured data;
selecting a first unsupervised machine learning model from a plurality of unsupervised machine learning models, one or more of the plurality of unsupervised machine learning models comprise different types of information associated with the claim as the claim progresses, and wherein selection of the first unsupervised machine learning model is based on a type of information associated with the first unsupervised machine learning model;
inputting the data associated with the claim into the selected first unsupervised machine learning model and receiving as output from the selected first unsupervised machine learning model an identification of a cluster of candidate claims to which the claim belongs;
generating a score corresponding to a predicted performance of the provider based on the predicted performance of the claim and the identification of the cluster of candidate claims; and
providing, for display at a client device, the generated score of the provider.
25 . The method of claim 24 , wherein one or more of the plurality of unsupervised machine learning models comprises a different complexity, and wherein selection of the first unsupervised machine learning model is further based on a complexity of the first unsupervised machine learning model.
26 . The method of claim 24 , wherein the type of information associated with the first unsupervised machine learning model comprises a type of specialty, a location of practice, a number of years a provider has been in practice, a type of service provided, a type of patient treated, or a type of insurance accepted, a type of claim handled, or combinations thereof.
27 . The method of claim 24 , wherein the different types of information are weighted based on a stage of claim processing associated with each machine learning model of the plurality of unsupervised machine learning models.