Interactive multimedia and methods for designing interactive multimedia based on decision trees generated by machine learning

1 . A non-transitory computer-readable medium having stored thereon a set of instructions that are executable by a processor of a computer system to carry out an interactive multimedia with multiple levels, to enable interactions with a user, wherein the interactive multimedia comprises:

a first level program, providing at least one first level interactive activity to determine one of multiple lines for the user to proceed from a first level to a next level; and

a second level program, providing at least one second level interactive activity to determine one of multiple lines for the user to proceed to from a second level to a next level,

wherein the levels and the lines of the interactive multimedia are designed based on a decision tree generated by steps of:

preprocessing a source data, which comprises handling missing data;

running a decision tree algorithm on the source data, which includes predictor variables indicating attributes of respondents and a target variable indicating related outcomes of the respondents, to generate the decision tree;

evaluating the decision tree generated by the decision tree algorithm with one or more predetermined criteria; and

in response to the decision tree not satisfying the one or more predetermined criteria, modifying the decision tree and re-evaluating the refined decision tree until the one or more predetermined criteria are satisfied.

2 . The non-transitory computer-readable medium according to claim 1 , wherein the respondents are individuals in a sample population.

3 . The non-transitory computer-readable medium according to claim 1 , wherein the respondents are targeted entities.

4 . The non-transitory computer-readable medium according to claim 1 , wherein the lines correspond to branches of the decision tree, and the interactive activities correspond to nodes of the decision tree.

5 . The non-transitory computer-readable medium according to claim 4 , wherein one of the lines from each interactive activity is determined according to a splitting rule at the corresponding node of the decision tree.

6 . The non-transitory computer-readable medium according to claim 1 , wherein an (N-1)th level program of the interactive multimedia provides at least one (N-1)th level interactive activity to determine one of multiple lines for the user to proceed to an Nth level program, which provides Nth level interactive activities for presenting the outcomes.

7 . The non-transitory computer-readable medium according to claim 6 , wherein each of the Nth level interactive activities further provide analysis based on the decision tree for explaining cause of the one of the outcomes.

8 . The non-transitory computer-readable medium according to claim 6 , wherein at least one of the Nth level interactive activities further provides a plan for the user based on the decision tree, to guide the user for going through a different combination of the lines leading to another one of the outcomes.

9 . The non-transitory computer-readable medium according to claim 8 , wherein the plan provided by the at least one of the Nth level interactive activities comprises interactions with the user, and is adjustable according to reactions provided by the user.

10 . The non-transitory computer-readable medium according to claim 1 , wherein the interactive activities respectively comprise conducting a conversation with the user, and/or respectively comprise presenting an event to the user and sensing reaction of the user in response to the event.

11 . The non-transitory computer-readable medium according to claim 1 , wherein the interactive activities respectively comprise tracking static and/or dynamic postures of the user.

12 . The non-transitory computer-readable medium according to claim 1 , wherein the interactive activities respectively comprise tracking and/or recording interactions between the user and at least one another.

13 . The non-transitory computer-readable medium according to claim 1 , wherein the modifying the at least one decision tree comprises modifying or pruning the at least one decision tree based on feature importance and/or variance of inflation factor (VIF) of the predictor variables.

14 . The non-transitory computer-readable medium according to claim 1 , wherein generating the decision tree further comprises performing a preliminary evaluation on the source data, based on feature importance and/or VIF.

15 . The non-transitory computer-readable medium according to claim 1 , wherein the interactive multimedia is designed and generated partially based on hierarchical tree structures of the decision tree.

16 . The non-transitory computer-readable medium according to claim 1 , wherein the interactive multimedia is designed and generated based on combination of multiple hierarchical tree structures of the decision trees.

17 . A computer implemented method for designing an interactive multimedia, comprising:

generating a decision tree based on a source data by using a machine learning technique, wherein the decision tree has a hierarchical tree structure with multiple layers connected with one another by multiple branches; and

generating user interface of the interactive multimedia, wherein the user interface has multiple levels corresponding to the multiple layers of the decision tree, and adjacent levels are connected via multiple lines corresponding to the multiple branches of the decision tree, respectively;

wherein the generating the decision tree comprises:

preprocessing the source data, which comprises handling missing data;

running a decision tree algorithm on the source data including predictor variables indicating attributes of respondents and a target variable indicating related outcomes of the respondents to generate the decision tree;

evaluating the decision tree generated by the decision tree algorithm with one or more predetermined criteria; and

in response to the decision tree not satisfying the one or more predetermined criteria, modifying the decision tree and re-evaluating the refined decision tree until the one or more predetermined criteria are satisfied.

18 . The computer implemented method according to claim 17 , wherein the interactive multimedia is designed and generated partially based on hierarchical tree structures of the decision tree.

19 . The computer implemented method according to claim 18 , wherein the respondents are individuals in a sample population, or targeted entities.

20 . The computer implemented method according to claim 18 , wherein the layers of the decision tree respectively have at least one node from which at least two of the branches branch out, and each level of the user interface provides at least one interactive activity for a user based on the at least one node in one of the layers of the decision tree.

21 . The computer implemented method according to claim 20 , wherein the nodes of the decision tree include a root node and multiple internal nodes selected from the predictor variables in the source data.

22 . The computer implemented method according to claim 20 , wherein each interactive activity at one of the levels of the user interface is provided for acquiring an attribute of the user, which is corresponding to one of the predictor variables in the source data.

23 . The computer implemented method according to claim 20 , wherein a bottom layer of the decision tree has multiple ones of the node indicating the related outcomes of the respondents, and a bottom level of the user interface provides multiple ones of the interactive activity for presenting the related outcomes.

24 . The computer implemented method according to claim 23 , wherein the bottom level of the user interface further provides analysis based on the decision tree for explaining causes resulting the outcomes.

25 . The computer implemented method according to claim 23 , wherein the bottom level of the user interface further provides a plan to guide the user to another one of the outcomes.

26 . The computer implemented method according to claim 18 , wherein the modifying the decision tree comprises modifying or pruning the decision tree based on feature importance and/or variance of inflation factor (VIF) of the predictor variables.

27 . The computer implemented method according to claim 18 , wherein generating the decision tree further comprises performing a preliminary evaluation on the source data, based on feature importance and/or VIF.

28 . The computer implemented method according to claim 18 , wherein the interactive multimedia is designed and generated based on combination of multiple hierarchical tree structures of the decision trees.

29 . A non-transitory computer-readable medium having stored thereon a set of instructions that are executable by a processor of a computer system to carry out an interactive multimedia with multiple levels; to execute computer program instructions to enable interactions with a user, and comprising:

a top level program, providing at least one top level interactive activity to determine a specific one of multiple lines for the user to proceed from a top level to a next level; and

at least one internal level program, providing at least one internal level interactive activity for each internal level to determine one of multiple lines for the user to proceed from the first internal level directly or indirectly to a bottom level,

wherein the top level interactive activity, the at least one internal level interactive activity and the bottom level interactive activity of the interactive multimedia are designed respectively based on a root node of a top layer, an internal node of a first internal layer, and a leaf node of a bottom layer of a decision tree generated by a machine learning technique with steps of:

preprocessing a source data, which comprises handling missing data;

running a decision tree algorithm on the source data including predictor variables indicating attributes of respondents and a target variable indicating related outcomes of the respondents to generate the decision tree;

evaluating the decision tree generated by the decision tree algorithm with one or more predetermined criteria; and

in response to the decision tree not satisfying the one or more predetermined criteria, modifying the decision tree and re-evaluating the refined decision tree until the one or more predetermined criteria are satisfied.

30 . The non-transitory computer-readable medium according to claim 29 , wherein (1) the at least one internal level of the interactive multimedia has multiple internal levels including the first internal level and a last internal level, (2) at least one internal level interactive activity for each internal level determines one of multiple lines for the user to procced from the first internal level directly or indirectly to the last internal level, and (3) the last internal level interactive activity of the interactive multimedia is designed based on an internal node of the last internal layer of the decision tree.

31 . The non-transitory computer-readable medium according to claim 29 , wherein (1) the at least one internal level of the interactive multimedia has at least three internal levels including the first internal level, a second internal level, and a last internal level, (2) at least one internal level interactive activity for each internal level is provided to determine one of multiple lines for the user to procced from the first internal level to the second internal level and then directly or indirectly to the last internal level, and (3) the second internal level interactive activity and the last internal level interactive activity of the interactive multimedia are designed respectively based on an internal node of the second internal layer and an internal node of the last internal layer of the decision tree.

32 . A computer implemented method for designing an interactive multimedia, comprising:

generating a decision tree based on a source data by using a machine learning technique, wherein the decision tree has a hierarchical tree structure with multiple layers connected with one another by multiple branches; and

generating user interface of the interactive multimedia, wherein the user interface has a top level, at least one internal level, and a bottom level, and provides at least one interactive activity at each level to determine one of multiple lines to proceed from a top level, to a first internal level, and directly or indirectly to a bottom level;

wherein proceeding from the top level, to the first internal level, and directly or indirectly to the bottom level of the user interface corresponds to a path from a root node of a top layer, to a first internal node of a first internal layer, and directly or indirectly to a leaf node of a bottom layer of the decision tree, respectively; and

wherein generating the decision tree comprises:

preprocessing the source data, which comprises handling missing data;

running a decision tree algorithm on the source data including predictor variables indicating attributes of respondents and a target variable indicating related outcomes of the respondents to generate the decision tree;

evaluating the decision tree generated by the decision tree algorithm with one or more predetermined criteria; and

in response to the decision tree not satisfying the one or more predetermined criteria, modifying the decision tree and re-evaluating the refined decision tree until the one or more predetermined criteria are satisfied.

33 . The computer implemented method according to claim 32 , wherein (1) the at least one internal level of the user interface has multiple internal levels including the first internal level and a last internal level, (2) the user interface provides at least one internal level interactive activity for each internal level to determine one of multiple lines to procced from the first internal level directly or indirectly to the last internal level, and (3) the last internal level of the user interface is designed based on an internal node of the last internal layer of the decision tree.