System and method for reducing feature calculations

A computer programming product comprising instructions embedded in a non-transitory computer readable storage medium that, when executed by at least one hardware processor, configure the at least one hardware processor to:

• identify one or more electronic data properties used to generate one or more features used as input to one or more pertinent data analytic models, wherein the one or more features comprise electronic data that has been processed into a form for use by and input into the one or more pertinent electronic data analytic models, wherein the processing of the electronic data comprises at least one of a group consisting of:
• filtering, selecting, transforming, and combinations thereof, the electronic data;
• associate the one or more identified electronic data properties to one or more corresponding features to create a data property to feature mapping;
• load a plurality of incoming electronic data records;
• determine whether an incoming electronic data record of the plurality of incoming electronic data records is a previously processed electronic data record;
• store, in response to an incoming electronic record not being a previously processed electronic data record, the incoming record as a new electronic data record;
• use the data to feature mapping to prepare a list of new features to generate based upon the electronic data stored in the new electronic data record;
• generate one or more features for the stored new electronic data record based upon the prepared list of new features to generate;
• determine, in response to an incoming electronic data record being a previously processed electronic data record, whether the incoming electronic data record matches the previously processed electronic data record;
• discard, in response to the incoming electronic data record matching a previously processed electronic data record, one of the incoming electronic data record or the previously processed electronic data record;
• identify, in response to the incoming electronic data record not matching the previously processed electronic data record, one or more electronic data properties in the incoming electronic data record that have changed;
• determine, in response to identifying one or more electronic data properties in the incoming electronic data record that have changed, one or more corresponding features associated with the one or more identified electronic data properties in the incoming electronic data record that have changed by using the data property to feature mapping to generate a list of features to generate that corresponds only to the one or more identified electronic data properties in the incoming electronic data record that have changed;
• generate only the one or more corresponding features associated with the one or more identified electronic data properties in the incoming electronic data record that have changed thereby reducing an amount of feature calculations generated comprising generating only the features identified on the list of features to regenerate;
• update the previously processed electronic data record with the one or more electronic data properties that have changed or storing the incoming electronic data record;
• and
• input in and utilize by the one or more pertinent electronic data analytic models only the one or more corresponding features associated with the one or more identified electronic data properties in the incoming electronic data record that have changed to determine a probability for an entity in a particular field thereby reducing an amount of processing and calculations by an electronic data analytic model.

A computer-implemented system to reduce an amount of processing performed to update electronic data records comprising:

• a non-transitory memory storage device storing program instructions;
• and
• a hardware processor having circuitry and logic to execute said program instructions, wherein the hardware processor is in communication with said memory storage device and in response to executing said program instructions is configured to:
• identify one or more electronic data properties used to generate one or more features used as input to one or more pertinent data analytic models, wherein the one or more features comprise electronic data that has been processed into a form for use by and input into the one or more pertinent electronic data analytic models, wherein the processing of the electronic data comprises at least one of a group consisting of:
• filtering, selecting, transforming, and combinations thereof, the electronic data;
• associate the one or more identified electronic data properties to one or more corresponding features to create a data property to feature mapping;
• load a plurality of incoming electronic data records into the system;
• determine whether an incoming electronic data record of the plurality of incoming electronic data records is a previously processed electronic data record;
• store, in response to an incoming electronic data record not being a previously processed electronic data record, the incoming record as a new electronic data record;
• use the data property to feature mapping to prepare a list of new features to generate based upon the electronic data stored in the new electronic data record;
• generate one or more features for the stored new electronic data record based upon the prepared list of new features to generate;
• determine, in response to an incoming electronic data record being a previously processed electronic data record, whether the incoming electronic data record matches the previously processed electronic data record;
• discard, in response to the incoming electronic data record matching a previously processed electronic data record, one of the incoming electronic data record or the previously processed electronic data record;
• identify, in response to the incoming electronic data record not matching the previously processed electronic data record, one or more electronic data properties in the incoming electronic data record that have changed;
• determine, in response to identifying one or more electronic data properties in the incoming electronic data record that have changed, one or more corresponding features associated with the one or more identified electronic data properties in the incoming electronic data record that have changed by using the data property to mapping feature to generate a list of features to regenerate that corresponds only to the one or more identified electronic data properties in the incoming electronic data record that have changed;
• generate only the one or more corresponding features associated with the one or more identified electronic data properties in the incoming electronic data record that have changed thereby reducing an amount of feature calculations generated comprises generating only the features identified on the list of features to regenerate;
• update the previously processed electronic data record with the one or more electronic data properties that have changed or storing the incoming electronic data record;
• and
• input in and utilize by the one or more pertinent electronic data analytic models only the one or more corresponding features associated with the one or more identified electronic data properties in the incoming electronic data record that have changed to determine a probability for an entity in a particular field thereby reducing the amount of processing and calculations by an electronic data analytic model reruns.