Pattern Recognition Using NLP-Based Tokenizing and Clustering Models

A method, comprising:

• receiving, via a web service executing on a user device, an input primary interaction;
• accessing, from a memory storing a plurality of clusters indicating interaction group patterns, the plurality of clusters indicating interaction group patterns, wherein each of the plurality of clusters comprises a respective primary interaction and one or more secondary interactions linked to that primary interaction, and wherein the plurality of clusters are identified by a plurality of non-fungible tokens, respectively;
• determining an interaction type associated with the input primary interaction;
• determining a non-fungible token associated with the input primary interaction based on the interaction type;
• comparing the non-fungible token associated with the input primary interaction to the plurality of non-fungible tokens respectively identifying the plurality of clusters;
• selecting a first cluster from the plurality of clusters based on a match between the non-fungible token associated with the input primary interaction and a first non-fungible token identifying the first cluster;
• determining the one or more secondary interactions in the first cluster as linked to the input primary interaction;
• retrieving the one or more secondary interactions linked to the input primary interaction from the plurality of clusters;
• generating a recommended group of interactions comprising the input primary interaction and the one or more retrieved secondary interactions;
• and
• providing, via the web service, the recommended group of interactions and an indication that the one or more retrieved secondary interactions are linked to the input primary interaction.

A non-transitory computer-readable medium storing instructions that when executed by a processor cause the processor to:

• receive, via a web service executing on a user device, an input primary interaction;
• access, from a memory storing a plurality of clusters indicating interaction group patterns, the plurality of clusters indicating interaction group patterns, wherein each of the plurality of clusters comprises a respective primary interaction and one or more secondary interactions linked to that primary interaction, and wherein the plurality of clusters are identified by a plurality of non-fungible tokens, respectively;
• determine an interaction type associated with the input primary interaction;
• determine a non-fungible token associated with the input primary interaction based on the interaction type;
• compare the non-fungible token associated with the input primary interaction to the plurality of non-fungible tokens respectively identifying the plurality of clusters;
• select a first cluster from the plurality of clusters based on a match between the non-fungible token associated with the input primary interaction and a first non-fungible token identifying the first cluster;
• determine the one or more secondary interactions in the first cluster as linked to the input primary interaction;
• retrieve the one or more secondary interactions linked to the input primary interaction from the plurality of clusters;
• generate a recommended group of interactions comprising the input primary interaction and the one or more retrieved secondary interactions;
• and
• provide, via the web service, the recommended group of interactions and an indication that the one or more retrieved secondary interactions are linked to the input primary interaction.