Method for mortgage fraud detection

Abstract

A method of detection of fraud in a mortgage application: in a computer system, maintaining a database of sales prices of real properties in a geographic area where the property is located; obtaining a valuation history for the property; obtaining historical sales data for similar properties in the geographic area; computing price ratios using these valuation histories; computing a distortion index based on the price ratios, the distortion index indicating the likelihood of a fraudulent valuation.

1 . A method of detecting fraud during a real estate transaction, the method comprising:
 using a computer processor to:   receive an estimated value of a subject real property;   access a database of spatial data, the spatial data comprising real estate prices in a geographic area in which the subject real property is located, the geographic area comprising at least one of the same zip code, city, or county as the subject real property;   generate a first spatial variance by computing ratios corresponding to the accessed real estate prices;   generate a second spatial variance by computing ratios corresponding to the accessed real estate prices; and   compute a spatial distortion based on the first spatial variance and the second spatial variance to indicate a likelihood of fraud.   
     
     
         2 . The method of  claim 1 , wherein the spatial data is generated by using sales data for properties in the geographic area. 
     
     
         3 . The method of  claim 1 , wherein the spatial data is generated by using an automated valuation model for properties in the geographic area. 
     
     
         4 . The method of  claim 1 , wherein the spatial data is generated by using a combination of sales data and an automated valuation model for properties in the geographic area. 
     
     
         5 . The method of  claim 1 , wherein the spatial data is generated from real estate prices from previous years. 
     
     
         6 . The method of  claim 1 , wherein the spatial data comprises property characteristics that are shared between the subject real property and properties in the geographic area. 
     
     
         7 . The method of  claim 1 , wherein one of the spatial variances comprise a ratio of the estimated value of the subject real property and a median real estate price of real property in the same zip code. 
     
     
         8 . The method of  claim 1 , wherein the spatial distortion comprises a difference between two years of spatial variances of the subject real property. 
     
     
         9 . The method of  claim 8 , wherein the two years comprise the current year and a previous year having the largest spatial variance. 
     
     
         10 . A method of detecting fraud during a real estate transaction, the method comprising:
 using a computer processor to:   receive an estimated value of a subject real property;   access a database of spatial data, the spatial data comprising real estate prices in a geographic area in which the subject real property is located, the geographic area comprising at least one of the same zip code, city, or county as the subject real property;   generate a set of spatial variances by computing ratios between a plurality of real estate prices of the subject real property and a plurality of real estate prices of properties in the geographic area;   compute a spatial distortion based on the set of spatial variances; and   produce a distortion ratio score to indicate a likelihood of fraud based on the spatial distortion.   
     
     
         11 . The method of  claim 10 , further comprising accessing a set of temporal data. 
     
     
         12 . The method of  claim 11 , further comprising generating a set of temporal variances, computing a temporal distortion based on the set of temporal variances, and computing a total distortion by adding the temporal distortion to the spatial distortion. 
     
     
         13 . A system of detecting fraud during a real estate transaction, the method comprising:
 a computer processor; and   a memory storing program instructions, said program instructions when executed by the computer processor causes the computer processor to:
 receive an estimated value of a subject real property; 
 access a database of spatial data, the spatial data comprising real estate prices in a geographic area in which the subject real property is located, the geographic area comprising at least one of the same zip code, city, or county as the subject real property; 
 generate a set of spatial variances by computing ratios between a plurality of real estate prices of the subject real property and a plurality of real estate prices of properties in the geographic area; 
 compute a spatial distortion based on the set of spatial variances; and 
 produce a distortion ratio score to indicate a likelihood of fraud based on the spatial distortion. 
   
     
     
         14 . The system of  claim 13 , wherein the spatial data is generated by using a combination of sales data and an automated valuation model for properties in the geographic area. 
     
     
         15 . The system of  claim 13 , wherein the spatial data is generated from real estate prices from previous years. 
     
     
         16 . The system of  claim 13 , wherein the spatial data comprises property characteristics that are shared between the subject real property and properties in the geographic area. 
     
     
         17 . The system of  claim 13 , wherein one of the spatial variances comprise a ratio of the estimated value of the subject real property and a median real estate price of real property in the same zip code. 
     
     
         18 . The system of  claim 13 , wherein the spatial distortion comprises a difference between two years of spatial variances of the subject real property. 
     
     
         19 . The system of  claim 18 , wherein the two years comprise the current year and a previous year having the largest spatial variance. 
     
     
         20 . A method of detecting fraud during a real estate transaction, the method comprising:
 using a computer data processor to:   access a database of real property prices in a geographic area in which a subject real property is located;   using data from the database or data from a requestor to generate a temporal data set comprising a current yearly real property price for the subject real property and a set of past yearly real property prices for the subject real property;   generate from the database a spatial data set comprising a current yearly real property price for real property with similar characteristics as the subject real property and a set of past yearly real property prices for real property with similar characteristics as the subject real property;   generate a set of temporal variances;   generate a set of spatial variances;   compute a spatial distortion based on the set of spatial variances;   compute a temporal distortion based on the set of temporal variances; and   compute a total distortion by adding the temporal distortion to the spatial distortion.