System and method for classifying an alphanumeric candidate identified in an email message

A computing device, comprising:

• one or more processors;
• and
• a non-transitory computer-readable storage medium having a plurality of instructions stored thereon, which, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
• receiving a plurality of email messages, each of the plurality of email messages comprising text and a subject and being associated with a sender and a recipient;
• parsing each of the plurality of email messages to identify at least one alphanumeric candidate;
• for each particular alphanumeric candidate:
• associating the particular alphanumeric candidate with an originating email in which the particular alphanumeric candidate was identified,
• determining email specific features related to the particular alphanumeric candidate, the email specific features being features of the originating email and including a position in the originating email at which the particular alphanumeric candidate is located, and one or more words preceding the particular alphanumeric candidate,
• analyzing the plurality of email messages to determine recipient specific features pertaining to the particular alphanumeric candidate, recipient specific features being related to use of the particular alphanumeric candidate in other emails of the plurality of emails having a same recipient as the originating email, and
• analyzing the plurality of email messages to determine recipient agnostic features pertaining to the particular alphanumeric candidate, recipient agnostic features being related to use of the particular alphanumeric candidate in emails of the plurality of emails not having the same recipient as the originating email;
• clustering the alphanumeric candidates based on the email specific features, the recipient specific features, and the recipient agnostic features to generate a plurality of clusters, wherein the alphanumeric candidates in each particular cluster of the plurality of clusters are more similar to each other than to other alphanumeric candidates in other clusters;
• associating an alphanumeric candidate type with each of the plurality of clusters;
• training an alphanumeric candidate type classifier based on the plurality of clusters and the associated alphanumeric candidate types;
• and
• utilizing the alphanumeric candidate type classifier to determine the type of an unclassified alphanumeric candidate in a later received email message.

A non-transitory computer-readable storage medium having a plurality of instructions stored thereon, which, when executed by one or more processors, cause the one or more processors to perform operations comprising:

• receiving a plurality of email messages, each of the plurality of email messages comprising text and a subject and being associated with a sender and a recipient;
• parsing each of the plurality of email messages to identify at least one alphanumeric candidate;
• for each particular alphanumeric candidate:
• associating the particular alphanumeric candidate with an originating email in which the particular alphanumeric candidate was identified,
• determining email specific features related to the particular alphanumeric candidate, the email specific features being features of the originating email and including a position in the originating email at which the particular alphanumeric candidate is located, and one or more words preceding the particular alphanumeric candidate,
• analyzing the plurality of email messages to determine recipient specific features pertaining to the particular alphanumeric candidate, recipient specific features being related to use of the particular alphanumeric candidate in other emails of the plurality of emails having a same recipient as the originating email, and
• analyzing the plurality of email messages to determine recipient agnostic features pertaining to the particular alphanumeric candidate, recipient agnostic features being related to use of the particular alphanumeric candidate in emails of the plurality of emails not having the same recipient as the originating email;
• clustering the alphanumeric candidates based on the email specific features, the recipient specific features, and the recipient agnostic features to generate a plurality of clusters, wherein the alphanumeric candidates in each particular cluster of the plurality of clusters are more similar to each other than to other alphanumeric candidates in other clusters;
• associating an alphanumeric candidate type with each of the plurality of clusters;
• training an alphanumeric candidate type classifier based on the plurality of clusters and the associated alphanumeric candidate types;
• and
• utilizing the alphanumeric candidate type classifier to determine the type of an unclassified alphanumeric candidate in a later received email message.