Adaptable on-deployment learning platform for driver analysis output generation

A computer-implemented method comprising:

• training a risk model with a first dataset containing a plurality of data points;
• and
• deploying the risk model by:
• receiving, from one or more vehicle sensors, sensor data inputs;
• generating, for one or more of the sensor data inputs, a pattern deviation output corresponding to an output of the risk model and an assessment of risk associated with a given driver experience;
• clustering the pattern deviation outputs to maximize a ratio of inter-cluster variance to intra-cluster variance;
• training a long short term memory (LSTM) for each cluster;
• performing a consistency analysis to verify, based on the LSTM for each cluster, consistency of the pattern deviation outputs in each cluster;
• modifying, when the consistency of the pattern deviation outputs in each cluster is verified, the risk model to reflect pattern deviation outputs associated with verified consistency;
• and
• identifying, when the consistency of the pattern deviation outputs in each cluster is not verified, sensor data inputs associated with deviated patterns that do not correspond to the plurality of data points of the first dataset as a second dataset containing a plurality of new data points.

One or more non-transitory computer-readable media storing instructions that, when executed by a computing platform comprising at least one processor, a communication interface, and memory, cause the computing platform to:

• train a risk model with a first dataset containing a plurality of data points;
• and
• deploy the risk model by:
• receiving, from one or more vehicle sensors, sensor data inputs;
• generating, for one or more of the sensor data inputs, a pattern deviation output corresponding to an output of the risk model and an assessment of risk associated with a given driver experience;
• clustering the pattern deviation outputs to maximize a ratio of inter-cluster variance to intra-cluster variance;
• training a long short term memory (LSTM) for each cluster;
• performing a consistency analysis to verify, based on the LSTM for each cluster, consistency of the pattern deviation outputs in each cluster;
• modifying, when the consistency of the pattern deviation outputs in each cluster is verified, the risk model to reflect pattern deviation outputs associated with verified consistency;
• and
• identifying, when the consistency of the pattern deviation outputs in each cluster is not verified, sensor data inputs associated with deviated patterns that do not correspond to the plurality of data points of the first dataset as a second dataset containing a plurality of new data points.