Classification and prediction of online user behavior using HMM and LSTM

1. A system, comprising:

a memory that stores computer executable components; and

a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise:

a training component that:

trains, using a set of training data comprising interactions of users with digital mediums, a long short-term memory model employing at least one cross entropy loss function for a plurality of layers of the long short-term memory model to predict users next actions during the interactions with the digital mediums, and predict probabilities of conversion objectives being achieved with the users, and

trains, using the predicted users next actions, a Hidden Markov model to classify positions of the users along defined paths to the conversion objectives;

an obtaining component that obtains user behavior data from a digital medium, the user behavior data defining an interaction of a user with the digital medium;

a prediction component that determines, using the long short-term memory model, based on the user behavior data, a predicted next action of the user within the digital medium; and

a classification component that classifies, using the Hidden Markov model, based on the user behavior data and the predicted next action, a position of the user along a defined path to a conversion objective of the digital medium.

2. The system of claim 1 , wherein the conversion objective comprises obtaining the user as a customer, achieving a financial transaction, or presentation of a communication.

3. The system of claim 1 , wherein the prediction component further determines, using the long short-term memory model, based on the user behavior data, a probability of that the conversion objective will be achieved with the user.

4. The system of claim 1 , wherein the defined path comprises a plurality of stages, and wherein the prediction component further determines, using the long short-term memory model, based on the user behavior data, a probability of that a selected stage of the plurality of stages will be reached by the user.

5. The system of claim 1 , further comprising:

an output component that guides the user to the conversion objective by providing a communication to the user based on a pattern of recent user behavior defined by the user behavior data.

6. The system of claim 1 , further comprising:

an output component that recommends or provides a digital medium interface to the user during use of the digital medium by the user.

7. A computer-implemented method, comprising:

training, by a system operatively coupled to a processor, using a set of training data comprising interactions of users with digital mediums, a long short-term memory model by employing at least one cross entropy loss function for a plurality of layers of the long short-term memory model to predict users next actions during the interactions with the digital mediums, and predict probabilities of conversion objectives being achieved with the users, and

training, by the system, the predicted users next actions, a Hidden Markov model to classify positions of the users along defined paths to the conversion objectives;

obtaining, by the system, user behavior data from a digital medium, the user behavior data defining an interaction of a user with the digital medium;

determining, by the system, using the long short-term memory model, based on the user behavior data, a predicted next action of the user within the digital medium; and

classifying, by the system, using the Hidden Markov model, based on the user behavior data and the predicted next action, a position of the user along a defined path to a conversion objective of the digital medium.

8. The computer-implemented method of claim 7 , wherein the conversion objective comprises obtaining the user as a customer, achieving a financial transaction, or presentation of a communication.

9. The computer-implemented method of claim 7 , further comprising:

determining, by the system, using the long short-term memory model, based on the user behavior data, a probability of that the conversion objective will be achieved with the user.

10. The computer-implemented method of claim 7 , wherein the defined path comprises a plurality of stages, and wherein the computer-implemented method further comprises determining, by the system, using the long short-term memory model, based on the user behavior data, a probability of that a selected stage of the plurality of stages will be reached by the user.

11. The computer-implemented method of claim 7 , further comprising:

guiding, by the system, the user to the conversion objective by providing a communication to the user based on a pattern of recent user behavior defined by the user behavior data.

12. A computer program product facilitating a process to guide a user at a digital medium, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:

train, by the processor, using a set of training data comprising interactions of users with digital mediums, a long short-term memory model by employing at least one cross entropy loss function for a plurality of layers of the long short-term memory model to predict users next actions during the interactions with the digital mediums, and predict probabilities of conversion objectives being achieved with the users, and

train, by the processor, using the predicted users next actions, a Hidden Markov model to classify positions of the users along defined paths to the conversion objectives;

obtain, by the processor, user behavior data from the digital medium, the user behavior data defining an interaction of the user with the digital medium;

determine, by the processor, using the long short-term memory model, based on the user behavior data, a predicted next action of the user within the digital medium; and

classify, by the processor, using the Hidden Markov model, based on the user behavior data and the predicted next action, a position of the user along a defined path to a conversion objective of the digital medium.

13. The computer program product of claim 12 ,

wherein the conversion objective comprises obtaining the user as a customer, achieving a financial transaction, or presentation of a communication.

14. The computer program product of claim 12 , wherein the program instructions are further executable by the processor to cause the processor to:

determine, by the processor, using the long short-term memory model, based on the user behavior data, a probability of that the conversion objective will be achieved with the user.

15. The computer program product of claim 12 ,

wherein the defined path comprises a plurality of stages, and

wherein the program instructions are further executable by the processor to cause the processor to determine, by the processor, using the long short-term memory model, based on the user behavior data, a probability of that a selected stage of the plurality of stages will be reached by the user.

16. The computer program product of claim 12 , wherein the program instructions are further executable by the processor to cause the processor to:

recommend or provide, by the processor, a communication or digital medium interface to the user during use of the digital medium by the user.

17. The computer program product of claim 12 , wherein the long short-term memory model comprises a first layer that employs pages of the digital medium as input, and a second layer associated with times of the interactions as input.

18. The system of claim 1 , wherein the long short-term memory model comprises a first layer that employs pages of the digital medium as input, and a second layer associated with times of the interactions as input.

19. The system of claim 18 , wherein the long short-term memory model comprises concatenate layer that combines a first output of the first layer with a second output of the second layer.

20. The computer-implemented method of claim 7 , wherein the long short-term memory model comprises a first layer that employs pages of the digital medium as input, and a second layer associated with times of the interactions as input.