Generating parameterized record from electronic messages

A method for generating parameterized records using communications data, the method comprising:

• encoding a plurality of electronic messages into a plurality of bag-of-words embeddings;
• inputting the plurality of bag-of-words embeddings into a machine learning model to identify one or more electronic messages containing one or more transaction records, wherein the machine learning model is trained to identify bag-of-words embeddings containing one or more transaction records, and wherein each transaction record of the one or more transaction records comprises corresponding item identifiers;
• transmitting, to a record identification server, a request to identify one or more server records associated with the one or more transaction records, wherein the request comprises the corresponding item identifiers, and wherein the record identification server uses the corresponding item identifiers to identify the one or more server records;
• determining, based on one or more mutable parameters of the one or more server records, a value for a period of time during which a user is allowed to perform an action associated with the one or more server records;
• receiving, from the record identification server, the one or more server records, wherein each server record of the one or more server records comprises corresponding parameters that correspond to a respective purchase record of a respective purchased item;
• and
• in response to determining that the corresponding parameters match one or more sets of a plurality of sets of criteria, generating a parameterized transaction record for the one or more transaction records, wherein the parameterized transaction record comprises the one or more transaction records and corresponding parameter data.

A non-transitory, computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:

• receiving, a plurality of electronic messages associated with a user;
• encoding the plurality of electronic messages into encoded vectors;
• inputting the encoded vectors into a machine learning model to identify one or more electronic messages containing one or more transaction records, wherein the machine learning model is trained to identify encoded vectors containing one or more transaction records, and wherein each transaction record of the one or more transaction records comprises corresponding item identifiers;
• transmitting, to a record identification server, a request to identify one or more server records associated with the one or more transaction records, wherein the request comprises the corresponding item identifiers, wherein the record identification server uses the corresponding item identifiers to identify the one or more server records, and wherein each server record corresponds to a transfer of an item associated with an item identifier;
• receiving, from the record identification server, the one or more server records, wherein each server record of the one or more server records comprises corresponding parameters that correspond to a respective purchase record of a respective purchased item;
• comparing, for each server record of the one or more server records, the corresponding parameters with a plurality of sets of criteria;
• and
• in response to determining that the corresponding parameters match one or more sets of the plurality of sets of criteria, generating a parameterized transaction record for the one or more server records, wherein the parameterized transaction record comprises the one or more transaction records and corresponding parameter data.