| CPC G06Q 20/4016 (2013.01) [G06N 5/025 (2013.01); G06N 20/00 (2019.01)] | 20 Claims |
|
1. A system, comprising:
a processor; and
a memory having stored thereon instructions that are executable by the processor to cause the system to perform operations comprising:
generating a graphical user interface (GUI) configured to display a proposed modification to a machine learning transaction classification rule, wherein the machine learning transaction classification rule is comprised of a plurality of data component values corresponding to a plurality of data factors;
performing an analysis operation on the machine learning transaction classification rule, wherein the analysis operation comprises:
accessing a first existing threshold value for a first data component value of the plurality of data component values, wherein values above the first existing threshold value for the first data component value indicate a first type of transaction classification, and wherein values below the first existing threshold value for the first data component value indicate a second different type of transaction classification;
determining an updated threshold value for the first existing threshold value, wherein determining the updated threshold value comprises:
accessing a machine learning transaction model that was created based on a training dataset that includes a plurality of labeled training examples corresponding to a plurality of previously conducted electronic transactions, wherein each of the plurality of labeled training examples are assigned a respective transaction classification from a plurality of transaction classifications, wherein the plurality of transaction classifications include the first type of transaction classification and the second type of transaction classification, and wherein the plurality of labeled training examples includes a first subset of transactions assigned to the first type of transaction classification and includes a second subset of transactions assigned to the second type of transaction classification;
determining, from the machine learning transaction model, that the updated threshold value provides a greater degree of accuracy in correctly classifying electronic transactions into one of at least the first type of transaction classification or the second type of transaction classification;
providing, to a user through the generated GUI, graphical indication information that includes an indication that the updated threshold value provides the greater degree of accuracy for the machine learning transaction classification rule;
receiving, from the user via the generated GUI, a selection indicating an acceptance of the updated threshold value; and
replacing, in the machine learning transaction classification rule, the first existing threshold value for the first data component value of the plurality of data component values with the updated threshold value.
|