Causal inference for optimized resource allocation

A system comprising:

• one or more processors;
• and
• at least one memory storing processor-executable instructions that, when executed by the one or more processors, cause the one or more processors to:
• receive, using a machine learning framework comprising a first machine learning model and a second machine learning model, historical data comprising a plurality of variables corresponding to (i) a plurality of actions and inactions with respect to a plurality of resource-requesting entities, and (ii) a plurality of outcomes corresponding to the plurality of actions and inactions;
• inputting, by the one or more processors, the historical data to the machine learning framework to receive one or more causal effect predictions, wherein:
• (a) the first machine learning model is configured to generate a predictive risk score associated with an event based at least in part on the historical data
• (b) the second machine learning model is configured to:
• (1) receive at least a portion of the historical data, the predictive risk score, and a directed acyclic graph data, and
• (2) generate the one or more causal effect predictions on an outcome of interest from an action taken on a given resource-requesting entity of the plurality of resource-requesting entities based at least in part on the predictive risk score, at least the portion of the historical data, and the directed acyclic graph data, and
• (c) training the second machine learning model comprises:
• (1) determining causal effect values for the plurality of resource-requesting entities based at least in part on the predictive risk score, at least the portion of the historical data, and the directed acyclic graph data,
• (2) apportioning the plurality of resource-requesting entities into a plurality of resource-requesting entity subgroups, and
• (3) ranking the plurality of resource-requesting entity subgroups by magnitude of the causal effect values, wherein the one or more causal effect predictions are based at least in part on the ranking of the plurality of resource-requesting entity subgroups;
• determine a subset of the plurality of resource-requesting entity subgroups based at least in part on the one or more causal effect predictions;
• and
• initiate performance of one or more prediction-based actions based at least in part on the subset of the plurality of resource-requesting entity subgroups.

One or more non-transitory computer-readable storage media including instructions that, when executed by one or more processors, cause the one or more processors to:

• receive, using a machine learning framework comprising a first machine learning model and a second machine learning model, historical data comprising a plurality of variables corresponding to (i) a plurality of actions and inactions with respect to a plurality of resource-requesting entities, and (ii) a plurality of outcomes corresponding to the plurality of actions and inactions;
• inputting, by the one or more processors, the historical data to the machine learning framework to receive one or more causal effect predictions, wherein:
• (a) the first machine learning model is configured to generate a predictive risk score associated with an event based at least in part on the historical data
• (b) the causal inference second machine learning model is configured to;
• (1) receive at least a portion of the historical data, the predictive risk score, and a directed acyclic graph data, and
• (2) generate the one or more causal effect predictions on an outcome of interest from an action taken on a given resource-requesting entity of the plurality of resource-requesting entities based at least in part on the predictive risk score, at least the portion of the historical data, and the directed acyclic graph data, and
• (c) training the second machine learning model comprises:
• (1) determining causal effect values for the plurality of resource-requesting entities based at least in part on the predictive risk score, at least the portion of the historical data, and the directed acyclic graph data,
• (2) apportioning the plurality of resource-requesting entities into a plurality of resource-requesting entity subgroups, and
• (3) ranking the plurality of resource-requesting entity subgroups by magnitude of the causal effect values, wherein the one or more causal effect predictions are based at least in part on the ranking of the plurality of resource-requesting entity subgroups;
• determine a subset of the plurality of resource-requesting entity subgroups based at least in part on the one or more causal effect predictions;
• and
• initiate performance of one or more prediction-based actions based at least in part on the subset of the plurality of resource-requesting entity subgroups.