Distributed architecture for enabling machine-learned event analysis on end user devices
A computer system develops models and generates decision logic based on the developed models. The decision logic is distributed to end user devices, and the end user devices are able to implement the decision logic to detect events, determine event sequences, and correlate the determined event sequences to predicted outcomes.
1. A computing system comprising:
one or more processors;
memory resources to store a set of instructions;
wherein the one or more processors access the instructions to:
associate an event library with a user, the event library defining a plurality of activities of the user;
develop a machine-learned model using the event library;
based on the machine-learned model, generate decision logic for the user;
provide the decision logic to an end user device without communicating the machine-learned model to the end user device, the decision logic correlating information determined from event sequences to an intent or interest of the user;
cause the end user device to perform operations that include:
detect, during an online session initiated by the user, the user performing one or multiple activities that correspond to events of the event library;
record the set of events in sequence to reflect an order in time in which each of the one or multiple activities take place;
implement the decision logic to determine a value representing an intent or interest of the user to perform a conversion event of making a purchase during the online session; and
implement a trigger if the value representing the intent or interest of the user indicates that the user is indecisive towards completing the conversion event.
2. The computer system of claim 1 wherein the one or more processors access the instructions to cause the end user device to perform operations that include: detecting a user activity that is a response to the trigger; and communicating data that is indicative of a user response to the trigger to the computer system over one or more networks.
3. The computer system of claim 2 , wherein the one or more processors access the instructions to:
receive the data that is indicative of the user response; and
train the machine-learning model using the received data from the end user device.
4. The computer system of claim 3 , wherein the one or more processors access the instructions to:
generate an updated decision logic using the machine-learning model, once the machine-learning model has been trained using the received data from the end user device; and
provide the updated decision logic to the end user device.
5. The computer system of claim 4 , wherein the one or more processors access the instructions to:
provide the updated decision logic to the end user device over one or more networks.
6. The computer system of claim 1 , wherein the decision logic causes the end user device to make a determination as to whether the user performed a predicted action as a response to the trigger being implemented.
7. The computer system of claim 1 , wherein the one or more processors access the instructions to cause the end user device to detect an activity the user performs through a service application associated with the computer system.
8. The computer system of claim 1 , wherein the one or more processors access the instructions to cause the end user device to detect an activity using a sensor that is located on the end user device.
9. The computer system of claim 1 , wherein the one or more processors access the instructions to cause the end user device to detect an activity that is performed using a third-party application or service.
10. The computer system of claim 1 , wherein the one or more processors access the instructions to implement the trigger by generating a communication of a particular type for the user.
11. The computer system of claim 10 , wherein the communication is an offer.