×

Feedback loop linked models for interface generation

  • US 9,881,340 B2
  • Filed: 05/16/2011
  • Issued: 01/30/2018
  • Est. Priority Date: 12/22/2006
  • Status: Active Grant
First Claim
Patent Images

1. A computer system configured to process request data employing at least first, second and third predictive models connected in a feedback loop configuration comprising:

  • an application server configured to generate web pages to serve as user interfaces on user systems for accepting data related to an insurance request from a user, and to receive, via a network, telematics data derived from sensors monitoring one or more of vehicles, property, goods and individuals;

    one or more computing devices, the computing devices comprising at least one central processing unit and a non-transitory computer readable medium, the non-transitory computer-readable medium storing computer-executable instructions;

    a computerized underwriting workflow predictive model, trained in accordance with prior data and outcomes, using data parameters selected, from a plurality of stored data parameters, by regression analysis, configured to dynamically select and output, automatically and without human intervention, one or more of a plurality of available underwriting workflow components, the available underwriting workflow components comprising conducting a loss control investigation, requesting underwriting review of an insurance policy, verifying the accepted data with documentary evidence, and automatically issuing or declining the insurance policy, to apply in underwriting the recommended insurance policy;

    a computerized pricing predictive model, trained in accordance with prior data and outcomes, using data parameters selected, from a plurality of stored data parameters, by regression analysis, configured to determine and output, automatically and without human intervention, a price corresponding to the recommended insurance policy determined based on the output from the computerized underwriting workflow predictive model, coverage parameters for the recommended insurance policy, and the received data, anda computerized coverage recommendation predictive model, trained in accordance with prior data and outcomes, using data parameters selected, from a plurality of stored data parameters, by regression analysis, configured to determine and output, automatically and without human intervention, the coverage parameters for the recommended insurance policy,wherein the predictive models are connected in a feedback loop configuration such that a price output from the pricing computerized predictive model is fed as an input to the computerized underwriting workflow predictive model and the computerized coverage recommendation predictive model, output from the computerized underwriting workflow predictive model is fed as an input to the computerized pricing predictive model and the computerized coverage recommendation predictive model, and an output from the computerized coverage recommendation predictive model is fed as an input to the computerized pricing predictive model and the computerized underwriting workflow predictive model,wherein the instructions cause the central processing unit to iteratively execute the computerized predictive models in feedback loops until one of;

    (a) outputs of the predictive models across iterations stabilize;

    (b) a variation in outputs exhibits no more than predetermined changes;

    or (c) a predetermined number of iterations have been completed;

    wherein the instructions further cause the central processing unit to, upon reaching one of (a), (b) or (c), provide output data based on the determination reached by executing the computerized predictive models in feedback loops to the application server, the output data comprising a plurality of recommended coverage sets;

    wherein the application server is further configured to generate for display an interactive user interface including data indicative of each one of the plurality of recommended coverage sets, fields for receipt of user selected parameter adjustments, the adjustments being output to the central processing unit by the application server, the central processing unit being configured to cause the computerized predictive models to re-execute, employing the user selected parameter adjustments as fixed or fuzzy parameters, to, automatically and without human intervention, (i) determine a plurality of updated recommended coverage sets, and return data indicative of the updated recommended coverage sets to the application server for display, or (ii) perform a predetermined number of iterations of processing by the computerized predictive models, and, responsive to no set of coverages being identified after performance of the predetermined number of iterations, return data indicative of a denial of request for coverages to the application server.

View all claims
  • 1 Assignment
Timeline View
Assignment View
    ×
    ×