Automated prediction of user response states based on traversal behavior

1. A method comprising:

receiving behavior data for multiple players of a computer-implemented game, the behavior data indicating interaction journeys by respective players within the game, each interaction journey comprising a sequence of actions selected from a predefined set of actions;

based on the behavior data, compiling behavior graph data in which interaction journeys are represented as respective directed graph structures in which each action in the corresponding sequence of actions is represented as a respective action node;

storing the behavior graph data in a graph database;

accessing label data that indicates assigned associations between a plurality of psychological labels and the behavior graph data, each psychological label pertaining to a respective psychological feature of player experience or player behavior;

using the behavior graph data and the associated label data, training an artificial neural network for automated label prediction, thereby providing a trained neural network; and

using the trained neural network, producing a psychological label prediction for a player based on a particular interaction journey by the player.

2. The method of claim 1 , further comprising:

extracting, from the behavior graph data, training data that comprises respective graph structures for a subset of the interaction journeys in the behavior data; and

assigning to each interaction journey in the training data at least one respective psychological label selected from a predefined set of psychological labels, thereby providing labeled training data, the training of the artificial neural network is performed using the labeled training data.

3. The method of claim 2 , wherein the artificial neural network comprises a neural network model (NN model), wherein said training operation comprises training the NN model for label prediction responsive to input of respective action sequences, thereby providing a trained NN model, said producing of the psychological label prediction being performed using the trained NN model.

4. The method of claim 2 , wherein each graph structure in the graph database models a single respective interaction session of an associated player with a game application.

5. The method of claim 3 , further comprising:

based on the behavior data, building an action embedding matrix that maps each respective action of the predefined set of actions to a corresponding action embedding vector;

using the action embedding matrix, converting a plurality of actions in a given action sequence to corresponding action embedding vectors; and

providing the resultant plurality of action embedding vectors as inputs to the NN model for the given action sequence at least one of:

the training of the NN model, and

label estimation for the given action sequence.

6. The method of claim 5 , wherein the building of the action embedding matrix comprises:

extracting multiple action sequences from the behavior data; and

using the multiple action sequences provided to an embedding NN model, optimizing a candidate embedding matrix to predict as output a terminal action of each action sequence responsive to input comprising a plurality of pre-terminal actions in the action sequence.

7. The method of claim 1 , wherein the artificial neural network is configured to produce a single respective psychological label prediction for each interaction journey received as input for label prediction.

8. The method of claim 1 , wherein one or more of the plurality of psychological labels pertains to player motivation for gameplay.

9. The method of claim 1 , wherein the artificial neural network is configured to produce a sequence of psychological label predictions for each respective interaction journey received as input for label prediction.

10. The method of claim 1 , wherein one or more of the plurality of psychological labels pertains to an emotional state of the player.

11. The method of claim 1 , further comprising:

incorporating the psychological label prediction into a parametric player model; and

in an automated procedure, generating custom game content based at least in part on the psychological label prediction incorporated in the parametric player model.

12. A system comprising:

one or more computer processor devices; and

memory storing instructions to configure the system, when the instructions are executed by the one or more computer processor devices, to perform operations comprising:

receiving behavior data for multiple players of a computer-implemented game, the behavior data indicating interaction journeys by respective players within the game, each interaction journey comprising a sequence of actions selected from a predefined set of actions;

based on the behavior data, compiling behavior graph data in which interaction journeys are represented as respective directed graph structures in which each action in the corresponding sequence of actions is represented as a respective action node;

storing the behavior graph data in a graph database;

accessing label data that indicates assigned associations between a plurality of psychological labels and the behavior graph data, each psychological label pertaining to a respective psychological feature of player experience or player behavior;

using the behavior graph data and the associated label data, training an artificial neural network for automated label prediction, thereby providing a trained neural network; and

using the trained neural network, producing a psychological label prediction for a player based on a particular interaction journey by the player.

13. The system of claim 12 , wherein the instructions further configure the one or more computer processor devices to perform operations comprising:

extracting, from the behavior graph data, training data that comprises respective graph structures for a subset of the interaction journeys in the behavior data; and

assigning to each interaction journey in the training data at least one respective psychological label selected from a predefined set of psychological labels, thereby providing labeled training data, the training of the artificial neural network is performed using the labeled training data.

14. The system of claim 13 , wherein the artificial neural network comprises a neural network model (NN model), and wherein said training operation comprises training the NN model for label prediction responsive to input of respective action sequences, thereby providing a trained NN model, said producing of the psychological label prediction being performed using the trained NN model.

15. The system of claim 14 , wherein the instructions further configure the one or more computer processor devices to perform operations comprising:

based on the behavior data, building an action embedding matrix that maps each respective action of the predefined set of actions to a corresponding action embedding vector;

using the action embedding matrix, converting a plurality of actions in a given action sequence to corresponding action embedding vectors; and

providing the resultant plurality of action embedding vectors as inputs to the NN model for the given action sequence at least one of:

the training of the NN model, and

label estimation for the given action sequence.

16. The system of claim 15 , wherein the building of the action embedding matrix comprises:

extracting multiple action sequences from the behavior data; and

using the multiple action sequences provided to an embedding NN model, optimizing a candidate embedding matrix to predict as output a terminal action of each action sequence responsive to input comprising a plurality of pre-terminal actions in the action sequence.

17. The system of claim 12 , wherein the artificial neural network is configured to produce a single respective psychological label prediction for each interaction journey received as input for label prediction.

18. The system of claim 12 , wherein the artificial neural network is configured to produce a sequence of psychological label predictions for each respective interaction journey received as input for label prediction.

19. The system of claim 12 , wherein the instructions further configure the system to, wherein the instructions further configure the apparatus to:

incorporating the psychological label prediction into a parametric player model; and

in an automated procedure, generating custom game content based at least in part on the psychological label prediction incorporated in the parametric player model.

20. A non-transitory computer-readable storage medium having stored thereon instructions that, when executed by one or more computer processor devices, cause the one or more computer processor devices to perform operations comprising:

receiving behavior data for multiple players of a computer-implemented game, the behavior data indicating interaction journeys by respective players within the game, each interaction journey comprising a sequence of actions selected from a predefined set of actions;

based on the behavior data, compiling behavior graph data in which interaction journeys are represented as respective directed graph structures in which each action in the corresponding sequence of actions is represented as a respective action node;

storing the behavior graph data in a graph database;

accessing label data that indicates assigned associations between a plurality of psychological labels and the behavior graph data, each psychological label pertaining to a respective psychological feature of player experience or player behavior;

using the behavior graph data and the associated label data, training an artificial neural network for automated label prediction, thereby providing a trained neural network; and

using the trained neural network, producing a psychological label prediction for a player based on a particular interaction journey by the player,

particular interaction journey by the player.