Intelligent data matching and validation system

Abstract

An intelligent data matching and validation system, and associated methods, are disclosed. The system includes various processors, databases, and immutable ledgers for analyzing data such as, but not limited to, loan data. The system leverages intelligent resources such as Bayesian inference networks for determining high correlation events representative of likely outcomes based on data parameters. The system automatically updates the Bayesian inference network's weighted coefficients in response to processing loan data and corresponding target events, such as repurchase requests. The system stores outcomes from the Bayesian inference networks, as well as loan histories and associated loan data, in a ledger that is accessible to third parties via a unique cryptographic token. In one embodiment, the system conducts an intelligent underwriting of a loan or other financial asset, leveraging its access to and ability to interpret and compare data from multiple sources that lead to the intelligent underwriting of a loan.

What is claimed is: 
     
         1 . A system comprising:
 a processor operatively connected to a server, wherein the processor is operatively configured to:
 generate a predictive data model by training the predictive data model with data from a plurality of historical financial loan applications, wherein the predictive data model is configured to generate one or more weighted coefficients corresponding to a Bayesian inference network; 
 in response to receiving one or more financial loan elements identified as unsatisfactory by one or more third parties, retrieve from the server corresponding loan data comprising information sampled from a financial loan application associated with the one or more financial loan elements identified as unsatisfactory; 
 compare the one or more financial loan elements identified as unsatisfactory to the retrieved corresponding loan data for determining if the one or more financial loan elements are unsatisfactory; and 
 automatically update the one or more weighted coefficients corresponding to the Bayesian inference network based on the comparison, wherein the one or more weighted coefficients are strengthened if the corresponding loan data falsifies the one or more financial loan elements identified as unsatisfactory, and wherein updating the one or more weighted coefficients comprises retraining the predictive data model with the updated one or more weighted coefficients. 
   
     
     
         2 . The system of  claim 1 , wherein the server comprises a blockchain ledger. 
     
     
         3 . The system of  claim 1 , wherein the server comprises a QLDB ledger. 
     
     
         4 . The system of  claim 1 , wherein the one or more financial loan elements received from the one or more third parties comprises data formatted in accordance with a Form  1003  mortgage application. 
     
     
         5 . The system of  claim 1 , wherein the corresponding loan data comprises underwriting history associated with the one or more financial loan elements. 
     
     
         6 . The system of  claim 1 , wherein the retrieved corresponding loan data comprises published investor guidelines, and wherein the published investor guidelines are published to the server by a particular third-party via an API providing permissioned access to the server via a web portal. 
     
     
         7 . The system of  claim 1 , wherein the retrieved corresponding loan data comprises loan data not directly related to the one or more financial loan elements received from the one or more third parties. 
     
     
         8 . The system of  claim 7 , wherein the loan data not directly related to the one or more financial loan elements received from the one or more third parties comprises sample data corresponding to loans similar to the one or more financial loan elements. 
     
     
         9 . The system of  claim 8 , wherein the processor retrieves the sample data according to an algorithm trained via deep learning techniques. 
     
     
         10 . A method, wherein via a processor operatively connected to a server, the processor is configured to execute the steps comprising:
 generating a predictive data model by training the predictive data model with data from a plurality of historical financial loan applications, wherein the predictive data model is configured to generate one or more weighted coefficients corresponding to a Bayesian inference network;   in response to receiving one or more financial loan elements identified as unsatisfactory by one or more third parties, retrieving from the server corresponding loan data comprising information sampled from a financial loan application associated with the one or more financial loan elements identified as unsatisfactory;   comparing the one or more financial loan elements identified as unsatisfactory to the retrieved corresponding loan data for determining if the one or more financial loan elements are unsatisfactory; and   automatically updating the one or more weighted coefficients corresponding to the Bayesian inference network based on the comparison, wherein the one or more weighted coefficients are strengthened if the corresponding loan data falsifies the one or more financial loan elements identified as unsatisfactory, and wherein updating the one or more weighted coefficients comprises retraining the predictive data model with the updated one or more weighted coefficients.   
     
     
         11 . The method of  claim 10 , wherein the server comprises a blockchain ledger. 
     
     
         12 . The method of  claim 10 , wherein the server comprises a QLDB ledger. 
     
     
         13 . The method of  claim 10 , wherein the one or more financial loan elements received from the one or more third parties comprises data formatted in accordance with a Form  1003  mortgage application. 
     
     
         14 . The method of  claim 10 , wherein the corresponding loan data comprises underwriting history associated with the one or more financial loan elements. 
     
     
         15 . The method of  claim 10 , wherein the retrieved corresponding loan data comprises published investor guidelines, and wherein the published investor guidelines are published to the server by a particular third-party via an API providing permissioned access to the server via a web portal. 
     
     
         16 . The method of  claim 10 , wherein the retrieved corresponding loan data comprises loan data not directly related to the one or more financial loan elements received from the one or more third parties. 
     
     
         17 . The method of  claim 16 , wherein the loan data not directly related to the one or more financial loan elements received from the one or more third parties comprises sample data corresponding to loans similar to the one or more financial loan elements. 
     
     
         18 . The method of  claim 17  wherein the processor retrieves the sample data according to an algorithm trained via deep learning techniques. 
     
     
         19 . A tangible, non-transitory, computer-readable medium comprising instructions encoded therein, wherein the instructions, when executed by one or more processors, cause the one or more processors to:
 generate a predictive data model by training the predictive data model with data from a plurality of historical financial loan applications, wherein the predictive data model is configured to generate one or more weighted coefficients corresponding to a Bayesian inference network;   in response to receiving one or more financial loan elements identified as unsatisfactory by one or more third parties, retrieve from the server corresponding loan data comprising information sampled from a financial loan application associated with the one or more financial loan elements identified as unsatisfactory;   compare the one or more financial loan elements identified as unsatisfactory to the retrieved corresponding loan data for determining if the one or more financial loan elements are unsatisfactory; and   automatically update the one or more weighted coefficients corresponding to the Bayesian inference network based on the comparison, wherein the one or more weighted coefficients are strengthened if the corresponding loan data falsifies the one or more financial loan elements identified as unsatisfactory, and wherein updating the one or more weighted coefficients comprises retraining the predictive data model with the updated one or more weighted coefficients.   
     
     
         20 . The tangible, non-transitory, computer-readable medium of  claim 19 , wherein the instructions, when executed by the one or more processors, further cause the one or more processors to generate the predictive data model wherein the predictive data model is configured to identify deviations in an individual's financial loan application, wherein the deviations indicate particular financial loan elements in the individual's financial loan application that are likely to comprise anomalies based on similar known unsatisfactory financial loan elements in the plurality of historical financial loan applications.