KR102654884B1 - Method for providing interactive information service and apparatus therefor - Google Patents

Abstract

The present invention relates to an interactive information provision service method and device for providing searched information to a user based on the user's selected response information to an interactive question. An information providing device according to the present invention includes a tree generator that generates a question tree based on attribute information defined for each question; When an interactive information provision service is requested from a user terminal, hierarchical questions formed in the question tree are sequentially provided to the user terminal, and the next level sub-question is selected based on the response information to the question and sent to the user terminal. A transmitting question provision unit; and an information providing unit that, when response information to the final question is received from the user terminal, extracts information corresponding to the response information to the final question and provides the information to the user terminal.

Description

{Method for providing interactive information service and apparatus therefor}

The present invention relates to an information provision service method, and more specifically, to an interactive information provision service method and apparatus for providing searched information to a user based on the user's selection response information to an interactive question.

Today, due to the development of communication devices and mobile communication networks, vast amounts of data exist on the Internet, and users often obtain desired information through search functions.

However, this passive search method varies depending on the search ability of each user. Users with good search ability may obtain the desired information immediately, while users with poor search ability may take a long time to obtain the desired information.

Meanwhile, technology has emerged to search for information within a specific category and provide it to users. The patent document below discloses a search method and system using a category identifier.

However, this category-based information search method also has a problem in that accurate information is provided to the user only when the user searches by entering an appropriate keyword, and inaccurate information is provided when the keyword is inappropriate. In addition, the conventional category-based information search method has the inconvenience of having to continuously change keywords when searching for inaccurate information.

Korean registered patent 10-0882437

The present invention was proposed to solve these conventional problems, and its purpose is to provide an interactive information provision service method and device that can provide users with easy-to-use and accurate search information.

Other objects and advantages of the present invention can be understood from the following description and will be more clearly understood by practicing the present invention. Additionally, it will be readily apparent that the objects and advantages of the present invention can be realized by the means and combinations thereof indicated in the patent claims.

An information providing device according to a first aspect of the present invention for achieving the above object includes a tree generator that generates a question tree based on attribute information defined for each question; When an interactive information provision service is requested from a user terminal, hierarchical questions formed in the question tree are sequentially provided to the user terminal, and the next level sub-question is selected based on the response information to the question and sent to the user terminal. A transmitting question provision unit; and an information providing unit that, when response information to the final question is received from the user terminal, extracts information corresponding to the response information to the final question and provides the information to the user terminal.

A method of performing an interactive information providing service in an information providing device according to a second aspect of the present invention for achieving the above object includes generating a question tree based on attribute information defined for each question; Receiving a request for an interactive information provision service from a user terminal; sequentially providing hierarchical questions formed in the question tree to the user terminal, and selecting a next-level sub-question based on response information to the question and transmitting it to the user terminal; And when response information to the final question is received from the user terminal, extracting information corresponding to the response information to the final question and providing it to the user terminal.

The present invention has the advantage of improving convenience when searching for information by extracting data based on the user's response information to a plurality of multiple-choice questions and providing it to the user.

In addition, the present invention creates a vivid question tree based on the attribute information of the question, sequentially provides highly correlated questions to the user based on the question tree, and provides information to the user based on the user's selected response to the question. Because it provides this, it has the effect of allowing users to accurately search for the information they require.

The following drawings attached to this specification illustrate preferred embodiments of the present invention, and serve to further understand the technical idea of the present invention along with specific details for carrying out the invention. Therefore, the present invention is described in such drawings. It should not be interpreted as limited to the specific details.
1 is a diagram illustrating a communication system according to an embodiment of the present invention.
Figure 2 is a diagram showing the configuration of an information provision device according to an embodiment of the present invention.
Figure 3 is a diagram illustrating a question including attribute information according to an embodiment of the present invention.
Figure 4 is a flowchart illustrating a method of generating a question tree in an information providing device, according to an embodiment of the present invention.
Figure 5 is a diagram illustrating the process of dividing questions into two clusters.
Figure 6 is a diagram illustrating the process of generating a question tree.
Figure 7 is a flowchart illustrating a method of providing an interactive information provision service to a user based on a question tree, according to an embodiment of the present invention.

The above-described purpose, features and advantages will become clearer through the following detailed description in conjunction with the accompanying drawings, and accordingly, those skilled in the art will be able to easily implement the technical idea of the present invention. There will be. Additionally, in describing the present invention, if it is determined that a detailed description of known technologies related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description will be omitted. Hereinafter, a preferred embodiment according to the present invention will be described in detail with reference to the attached drawings.

1 is a diagram illustrating a communication system according to an embodiment of the present invention.

As shown in FIG. 1, the communication system according to an embodiment of the present invention includes a user terminal 100 and an information provision device 200, and the user terminal 100 and the information provision device 200 are connected to a network. They communicate with each other through (300). The network 300 includes a mobile communication network and a wired Internet network, and since it corresponds to the well-known common technology of the present invention, detailed description will be omitted.

The user terminal 100 is a communication device owned by the user, and connects to the information provision device 200 to receive interactive information services. In particular, the user terminal 100 connects to the information provision device 200, sequentially receives multiple-choice questions, responds to each multiple-choice question, and receives a data group matching the response from the information provision device 200. do. The multiple-choice question is a multiple-choice type question, such as a multiple-choice question that includes a plurality of selection information, and an either-or question that allows one to choose between yes and no. In addition, a data group is information in which one or more pieces of data are collected, for example, an API data group in which APIs (application programming interfaces) are collected according to use, a department data group divided by department, a travel destination data group divided according to travel destination, etc. am. The user terminal 100 may be a mobile communication terminal such as a smartphone or a personal computer such as a desktop computer.

The information provision device 200 stores questions for which attribute information is defined and encoding information is assigned, and data groups corresponding to these questions, classified by data type. Additionally, the information provision device 200 uses a genetic algorithm to separate each question into optimal clusters and generates a question tree for each data type based on the separated clusters.

The genetic algorithm is a calculation model developed based on the natural evolutionary process and is one of the techniques used to derive the optimal solution. The genetic algorithm goes through selection, crossover, and mutation operations to derive the optimal solution to a specific problem. In an embodiment of the present invention, the information providing device 200 sets each question as a node and repeatedly performs selection operation, crossover operation, mutation operation, and fitness evaluation on the question using the encoding information assigned to each question. , each question is separated into a certain number of optimal clusters. Here, the encoding information represents a certain number of digits of identification information assigned to each question, corresponds to genetic information (e.g., chromosome information) in the genetic algorithm, and is generated in advance and assigned to each question. A certain number of digits corresponding to a decimal number are recorded in the encoding information. Optionally, data in various formats such as binary, octal, hexadecimal, etc. may be recorded in the encoding information.

The information providing device 200 provides a question to the user terminal 100, continues to provide another related question based on the response information selected by the user terminal 100, and then provides response information for the final question. Information corresponding to is extracted and provided to the user terminal 100. The information providing device 200 may provide any one of an API data group, a department data group divided by department, a travel destination data group divided according to a travel destination, etc. to the user terminal 100 as information.

Figure 2 is a diagram showing the configuration of an information provision device according to an embodiment of the present invention.

As shown in FIG. 2, the information providing device 200 according to an embodiment of the present invention includes a question providing unit 210, a storage unit 220, a tree generating unit 230, and an information providing unit 240. These components may be implemented as hardware or software, or through a combination of hardware and software.

The question provider 210 performs a function of providing a question containing a plurality of options and questions to the user terminal 100 based on the question tree formed in the tree generator 230. The question provider 210 provides a question corresponding to the root load as an initial question to the user terminal 100. In addition, the question provider 210 extracts a question of a child node corresponding to the selection response information from among the child nodes of the current question node, based on the selection response information of the user terminal 100, and provides the question to the user terminal 100. do.

The storage unit 220 is a storage medium such as memory, a disk device, etc., and stores a plurality of optional questions for each data type and also stores information (i.e., data group) corresponding to the questions. Additionally, the storage unit 220 stores question trees for each data type. In the question, attribute information is defined, encoding information is assigned, and a plurality of choice information is included in the question. Additionally, information on the relevance of each data group is recorded in the attribute information. The attribute information is used to calculate the suitability of cluster separation.

Figure 3 is a diagram illustrating a question including attribute information according to an embodiment of the present invention.

If explained as an example with reference to FIG. 3, the question includes attribute information in the form of multidimensional coordinates. For example, the question “Is it related to land transportation?” includes attribute information of (1,1,1,1,1,1,1,1,1,0,0,0,0), and “It is related to maritime transportation.” The question ‘Is it there?’ includes attribute information of (0,0,0,0,0,0,0,0,0,1,0,0,0). Each attribute information is expressed as a multidimensional coordinate value. and '0' or '1' is recorded as a detailed value. Here, '1' is flag information indicating that there is a relationship with the data group, and '0' is flag information indicating that there is no relationship with the data group. According to the attribute information included in the question “Is it related to maritime traffic?” in FIG. 3, the question is related to the “ship operation” data group and is not related to any other data group.

The tree generator 230 uses a genetic algorithm to separate each question into a certain number of optimal clusters, generates a question tree for each data type based on the separated clusters, and stores it in the storage unit 220. That is, the tree generator 230 uses a genetic algorithm to generate a question tree with a hierarchical structure according to the correlation of questions and stores it in the storage unit 220.

Specifically, the tree generator 230 initially separates a plurality of questions into a certain number of optimal clusters based on factors such as question attribute information and question encoding information. At this time, the tree generator 230 separates the questions into a certain number of arbitrary clusters, calculates the internal similarity and external similarity for the clusters using Equations 1 and 2, which will be described later, and the internal similarity and internal similarity are After the multiplied final similarity is calculated, the selection operation, crossover operation, mutation operation, and fitness evaluation (i.e., final similarity convergence evaluation) are repeated until the final similarity converges to initially separate the questions into optimal clusters.

The selection operation selects a question that is subject to crossover among questions, and an equal proportional roulette wheel selection technique, tournament selection technique, rank-based selection technique, etc. may be used. The crossover operation is a process of creating a new object (i.e., encoding information) by crossing the selected questions, and the new object created by crossing inherits the parent's genes (i.e., part of the encoding information). In the crossing operation, cycle crossing operation, order crossing operation, arithmetic crossing operation, heuristic crossing operation, etc. may be used. The mutation operation represents the process of deriving a mutant object by changing some data of the encoding information of a newly created object into data (i.e., numbers) that are not inherited from the parent (i.e., the question selected in the selection operation). . The data to be changed may be generated through a random number generator.

The object for which the mutation operation has been completed (i.e., encoding information) is changed to the encoding information of a specific questionnaire. At this time, the tree creation unit 230 may check the question that has encoding information most similar to the object for which the mutation operation has been completed, and change the encoding information of this question to the encoding information for which the mutation operation has been completed.

When one cycle according to the selection operation, crossover operation, and mutation operation is completed, the encoding information of the specific question is updated. Then, the tree generator 230 substitutes the attribute information of the question belonging to the currently separated cluster into Equation 1 and Equation 2, calculates the final similarity of the currently separated cluster, and determines whether this similarity is converging. If convergence is achieved, the currently separated cluster is determined to be the optimal cluster and cluster separation is stopped.

When the optimal cluster is initially separated, the tree generator 230 generates a question with attribute information corresponding to a distance close to the center value of the first cluster as the root load. Additionally, the tree creation unit 230 generates a specific question included in the initially separated second cluster as a second child node of the root load. At this time, the tree generator 230 may generate a question with attribute information corresponding to the closest distance to the center value of the second cluster as the second child node of the root load.

In addition, the tree generator 230 separates each of the separated first and second clusters into a certain number of sub-clusters through the selection operation, crossover operation, and mutation operation of the genetic algorithm. Likewise, the tree generator 230 separates the optimally separated clusters into arbitrary sub-clusters, calculates the internal similarity and external similarity of the sub-clusters using Equations 1 and 2, and then calculates the internal similarity and external similarity. By calculating this multiplied final similarity and repeating cluster separation through the genetic algorithm until the final similarity converges, the separated clusters are again separated into optimal sub-clusters.

The tree generator 230 generates representative questions of the sub-cluster with the highest frequency of the relevance flag (i.e., '1') among the sub-clusters of the first cluster (i.e., attribute information corresponding to the closest distance to the center value of the sub-cluster). A question with ) is created as the first child node (e.g., right child node) of the root node. In addition, the tree generator 230 generates representative questions of the sub-cluster with the lowest frequency of the relevance flag (i.e., '0') among the sub-clusters of the first cluster (i.e., corresponding to the closest distance to the center value of the sub-cluster). question with attribute information) is created as a child node of the first child node. In addition, the tree generator 230 asks the representative question of the sub-cluster with the highest frequency of the relevance flag (i.e., '1') among the sub-clusters of the second cluster to the child node of the second child node (e.g., the right child node). , and a representative question from a sub-cluster with a low frequency of relevance flag (i.e., '0') among the sub-clusters of the first cluster is created as another child node of the second child node.

While repeating the process of dividing each question into sub-clusters until each question is created as a node in the question tree, the tree creation unit 230 continues the process of creating child nodes for the parent node.

Meanwhile, the number of child nodes is created in correspondence with the number of selection information of the question. In other words, in the case of a question with alternative selection information such as 'yes' and 'no', a maximum of two child nodes can be created, and in the case of a question with three selection information, a maximum of three child nodes can be created. It can be.

When response information to the last question is received from the user terminal 100, the information provider 240 extracts information (i.e., a data group) corresponding to the last question and response information to the storage unit 220, Information is provided to the user terminal 100.

Figure 4 is a flowchart illustrating a method of generating a question tree in an information providing device, according to an embodiment of the present invention.

Figure 5 is a diagram illustrating the process of dividing questions into two clusters.

Figure 6 is a diagram illustrating the process of generating a question tree.

Hereinafter, in the description referring to FIGS. 4 and 6, it is assumed that the question is an either-or question and a binary tree is formed accordingly. In addition, questions with high relevance to the parent node are created in the left child node (i.e., child nodes corresponding to positive responses), and questions with low relevance are created in the right child node (i.e., child nodes corresponding to negative responses). It is assumed that

Referring to FIGS. 4 to 6, the tree generator 230 initially separates a plurality of questions into two optimal clusters through the selection operation, crossover operation, mutation operation, and fitness evaluation of the genetic algorithm (S401). The tree generator 230 uses the encoding information assigned to each question to perform a crossover operation and mutation operation for each question. In addition, the tree generator 230 uses a genetic algorithm to separate the questions into two random clusters, and calculates the internal similarity and external similarity of the randomly separated clusters using Equations 1 and 2 below, Cluster separation is repeatedly performed until the final similarity multiplied by the internal similarity and external similarity converges, and the questions are initially separated into optimal bipartite clusters. In detail, the tree generator 230 uses the encoding information assigned to each question as genetic information to perform selection, crossover, and mutation operations to separate clusters, and uses the final similarity as a fitness evaluation index.

Here, Fit _intra means internal similarity, k is the number of clusters, C _k is the center value of the k th cluster, and X _i represents the attribute information (i.e. multidimensional coordinate value) of the ith question belonging to the same cluster. .

Here, Fit _inter means external similarity, k _is the number of clusters, C _k is the center value of the k th cluster, d means the maximum k value.

Next, when a plurality of questions are initially separated into optimal bipartite clusters, the tree creation unit 230 selects one of the initially separated clusters A and B as the root target cluster. At this time, the tree generator 230 may select a cluster with a high frequency of association flag information '1' (i.e., total number of 1s) in the attribute information as the root target cluster, and conversely, the cluster with a high frequency of association flag information '0' may be selected as the root target cluster. A cluster with a small number can be selected as the root target cluster. In the explanation referring to FIGS. 4 to 6, it is explained that cluster A is selected as the root target cluster.

Figure 5 is a diagram that simplifies multidimensional questions into two dimensions, meaning that the closer to the left relative to the X coordinate axis, the more frequent the related flag information '1' is. For example, with reference to FIG. 5, the tree generator 230 forms questions 1 to 4 into cluster A, for a total of 7 questions, and questions 5 to 7. By forming cluster B, the questions are separated into two clusters, and among these, cluster A can be selected as the root target cluster.

Next, the tree generator 230 checks the center value of the selected root target cluster (i.e., cluster A) and asks a question with attribute information (i.e., multidimensional coordinate information) corresponding to the closest distance to this center value. Created as a root load (S403). Referring to Figure 5 (a) and Figure 6 (b), question number 3, which is closest to the center value of cluster A, the root target cluster, is generated as the root load.

Next, the tree generator 230 checks the center value of cluster B, which was not selected as the root target cluster, and generates a question with attribute information closest to the center value of this cluster as the first right child node of the root node. Do it (S405). Here, the right child node is a node with low correlation with the parent node and is a question node corresponding to a negative response. Referring to (a) of Figure 5 and (a) of Figure 6, question number 7, which is the closest to the center value of cluster B, which was not selected as the root target cluster, is created as the right child node of the root load.

Next, the tree creation unit 230 uses a genetic algorithm (i.e., selection operation, crossover operation, mutation operation) to separate each initially separated cluster into two optimal sub-clusters (S407). At this time, the tree generator 230 separates the initially separated cluster into two sub-clusters and calculates the internal similarity and external similarity of the sub-clusters using Equations 1 and 2 above. Then, the tree generator 230 calculates the final similarity by multiplying the internal similarity and the external similarity, and repeats the process of re-executing cluster separation through the genetic algorithm until the final similarity converges, thereby optimally dividing the separated clusters. Separate each into subclusters.

Next, the tree creation unit 230 proceeds with a process of generating questions included in the separated sub-cluster as child nodes (S409). That is, the tree generator 230 identifies the subcluster with the highest frequency of the relevance flag (i.e., '1') among the subclusters (i.e., A-1 cluster and A-2 cluster) included in the root target cluster, and , a representative question of this sub-cluster (i.e., a question with attribute information corresponding to the closest distance to the center value of the sub-cluster) is created as the first left child node of the root node. In addition, among the sub-clusters included in the root target cluster, representative questions of sub-clusters with a low frequency of relevance flags (i.e., '0') (i.e., questions with attribute information corresponding to the closest distance to the center value of the sub-cluster) is created as the right child node of the first left child node. At this time, the tree creation unit 230 selects and creates child nodes from questions that have not been selected as nodes.

Next, the tree generator 230 presents a representative question of the sub-cluster with the highest frequency of the relevance flag (i.e., '1') among the sub-clusters of cluster B (i.e., cluster B-1 and cluster B-2). 1Created as a left child node of the right child node (i.e., child node to the right of the root node), and a representative question of a subcluster with a low frequency of relevance flag (i.e., '0') among the subclusters of cluster B is asked in the first right child node. Created as the right child node of the child node.

Referring to Figure 5(b) and Figure 6(b), cluster A is separated into cluster A-1 and cluster A-2, and cluster B is separated into cluster B-1 and cluster B-2. And among clusters A-1 and A-2, question 2, which is the node closest to the center of cluster A-1, which is the subcluster with the highest frequency of relevance flags, is the left child node of the root node (i.e., question 3). is created with In addition, question 4, which is a node close to the center of the A-2 cluster, the subcluster with the lowest frequency of relevance flags, is created as the right child node of question 2. Additionally, among clusters B-1 and B-2, question 5, which is the node closest to the center of cluster B-1, which is the subcluster with the highest frequency of relevance flags, is the right child node of the root node (i.e., question 7). ) is created as the left child node of ). In addition, question 6, which is a node close to the center of the B-2 cluster, the subcluster with the lowest frequency of relevance flags, is created as the right child node of the root node's right child node (i.e., question 7).

Next, the tree creation unit 230 determines whether a node has been formed in the tree in response to the termination condition and the number of all questions (S411). If the termination condition is not met, re-proceeds step S407 to create another child node. Create. Referring to (c) of FIG. 5 and (c) of FIG. 6, the tree creation unit 230 re-proceeds from step S407 and divides the A-1 cluster into two sub-clusters (i.e., A-1-1, A-1 -2), and among the A-1-1 cluster and A-1-2 cluster, the question close to the center value of A-1-1 with the highest frequency of association flags is formed as the left child node of question 2.

Through this iterative process, a question tree is formed. The tree creation unit 230 receives a data type (e.g., API programming, company, etc.) from the administrator, stores the input data type and the formed question tree together in the storage unit 220, and stores the input data type and the formed question tree together in the storage unit 220. Question trees for each data type are stored in 220.

Meanwhile, in the above-described embodiment, the question tree is described as a binary tree, but a question tree having three or more child nodes can be formed. For example, the tree creation unit 230 can generate up to three child nodes for a plurality of questions having three types of selection information. In this case, the tree generator 230 selects the question with the highest correlation with the parent node as the leftmost (or rightmost) child node, the question with the second highest correlation with the parent node as the middle child node, and the question with the highest correlation with the parent node. The lowest question can be created as the rightmost (or leftmost) child node, respectively.

Figure 7 is a flowchart illustrating a method of providing an interactive information provision service to a user based on a question tree, according to an embodiment of the present invention.

Referring to FIG. 7, the question provider 210 receives a request for an interactive service from the user terminal 100. Then, the question provider 210 provides a data type list to the user terminal 100 and receives the specific data type selected by the user from the user terminal 100 (S701).

Next, the question provider 210 checks the question tree corresponding to the specific data type in the storage unit 220 (S703), and sends a question corresponding to the highest level (i.e., root) in this question tree to the user terminal. Send to (100) (S705). The question includes optional information.

Next, the question provider 210 receives selective response information for the question from the user terminal 100 (S707), checks the child node of the next level (i.e., level 2) corresponding to the selective response information, and , a question corresponding to this child node is provided to the user terminal 100 (S709). If the questionnaire is an either-or question and the selection response information is a positive response, the question provider 210 may provide the user terminal 100 with a question corresponding to the left (or right) child node of the root load. As another example, when the questionnaire is an either-or question and the selection response information is a negative response, the question provider 210 provides the user terminal 100 with a question corresponding to the right (or left) child node of the root load. can do. As another example, when the questionnaire is a multiple-choice question and the selected response information is answer 1, the question provider 210 may provide a question corresponding to the child node located on the far left to the user terminal 100. , if the selection response information is response number 2, the child node located second from the left can be provided to the user terminal 100.

Next, the question provider 210 receives selection response information for the question corresponding to the child node from the user terminal 100 (S711). Then, the question provider 210 determines whether the question has been completed (S713), and if so, instructs the information provider 240 to provide data.

Then, the information provider 240 extracts information (i.e., data group) corresponding to the last question and the selection response information to the storage unit 220, and stores this information (i.e., data group) in the user terminal 100. Provided as (S715).

*On the other hand, if it is determined that the question has not been completed in step S713, the question provider 210 checks the child node of the next level corresponding to the selection response information and sends the question corresponding to this child node to the user terminal ( Re-proceed step S709 provided by 100).

As described above, the present invention improves convenience when searching for information by extracting information and providing it to the user based on the user's response information to a plurality of multiple-choice questions. In addition, the present invention creates a vivid question tree based on the attribute information of the question, sequentially provides relevant questions to the user based on the question tree, and provides information to the user based on the user's selected response to the question. Because of this, users can accurately search for the information they require.

While this specification includes many features, such features should not be construed as limiting the scope of the invention or the scope of the claims. Additionally, features described in individual embodiments herein may be combined and implemented in a single embodiment. Conversely, various features described in a single embodiment herein may be implemented individually or in combination as appropriate in various embodiments.

Although operations in the drawings are illustrated in a particular order, it should not be understood that such operations are performed in a particular order as shown, or that they are performed in a sequential order, or that all of the described operations are performed to achieve a desired result. . Multitasking and parallel processing can be advantageous in certain environments. In addition, it should be understood that the division of various system components in the above-described embodiments does not require such division in all embodiments. The program components and systems described above may generally be implemented as a package in a single software product or multiple software products.

The method of the present invention as described above can be implemented as a program and stored in a computer-readable form on a recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.). Since this process can be easily performed by a person skilled in the art to which the present invention pertains, it will not be described in further detail.

The present invention described above is capable of various substitutions, modifications and changes without departing from the technical spirit of the present invention to those of ordinary skill in the technical field to which the present invention pertains. It is not limited by the drawings.

100: User terminal 200: Information provision device
210: question provision unit 220: storage unit
230: tree creation unit 240: information provision unit
300: Network

Claims (20)

Separating a plurality of questions into a plurality of clusters;
Generating a question tree in which hierarchical questions are formed based on the plurality of separated clusters; and
Providing hierarchical questions formed in the question tree sequentially to the user, and selecting a question of a lower node based on response information to the question,
The step of generating the question tree is,
Generating the question tree so that a plurality of questions included in different clusters are created as a plurality of child nodes of the same parent node; comprising:
How to provide information. According to clause 1,
The step of separating the plurality of questions into a plurality of clusters is:
Separating the plurality of questions into a certain number of the plurality of clusters by performing a genetic algorithm including at least one of a selection operation, a crossover operation, and a mutation operation.
How to provide information. According to clause 1,
The step of separating the plurality of questions into a plurality of clusters is:
Separating the plurality of questions into a certain number of clusters using attribute information assigned to the plurality of questions,
The attribute information includes information about the correlation between the question and a plurality of data groups.
How to provide information. According to clause 2,
The step of separating the plurality of questions into a plurality of clusters is:
For the cluster currently separated by the genetic algorithm, the final similarity of the currently separated cluster is calculated using the attribute information of the question belonging to the currently separated cluster, and the currently separated cluster is determined based on the final similarity. Including a step of determining that it is an optimal cluster.
How to provide information. delete According to clause 1,
The step of generating the question tree so that a plurality of questions included in the different clusters are generated as a plurality of child nodes of the same parent node,
A first question included in a first cluster among the plurality of clusters is generated as a first child node of the parent node, and a second question included in a second cluster different from the first cluster among the plurality of clusters is generated as the parent node. Created as a second child node of a node
How to provide information. According to clause 6,
The parent node is,
Generated from questions included in the first cluster
How to provide information. According to clause 7,
The step of generating the first question included in the first cluster among the plurality of clusters as the first child node of the parent node,
Separating questions included in the first cluster into a plurality of sub-clusters; and
Generating the first question included in a first sub-cluster among the plurality of sub-clusters as the first child node of the parent node; comprising
How to provide information. According to clause 8,
The first sub-cluster is,
The subcluster with the highest frequency of relevance flag among the plurality of subclusters
How to provide information. According to clause 8,
The step of generating the first question included in the first cluster among the plurality of clusters as the first child node of the parent node,
Generating a third question included in a second sub-cluster among the plurality of sub-clusters as a child node of the first child node,
The second sub-cluster is,
The subcluster with the lowest frequency of relevance flag among the plurality of subclusters
How to provide information. According to clause 1,
The step of generating the question tree so that a plurality of questions included in the different clusters are generated as a plurality of child nodes of the same parent node,
A step of generating a representative question of the cluster as a node,
The representative question above is,
A question with attribute information corresponding to the closest distance to the center value of the cluster.
How to provide information. According to clause 1,
The step of generating the question tree so that a plurality of questions included in different clusters are created as a plurality of child nodes of the same parent node,
A step of determining the location of a child node based on whether the relationship with the parent node is high or low,
The relevance to the parent node is determined based on the frequency of the relevance flag.
How to provide information. According to clause 1,
The plurality of clusters is,
Initially separated clusters, or sub-clusters separated from one of the initially separated clusters
How to provide information. a storage unit that stores a plurality of questions;
a tree generator that separates the plurality of questions into a plurality of clusters and generates a question tree in which hierarchical questions are formed based on the separated plurality of clusters; and
A question provider sequentially provides hierarchical questions formed in the question tree to the user, and selects questions of lower nodes based on response information to the questions,
The tree creation unit,
Generating the question tree so that multiple questions included in different clusters are created as multiple child nodes of the same parent node.
Information provision device. According to clause 14,
The tree creation unit,
Separating the plurality of questions into a certain number of the plurality of clusters by performing a genetic algorithm including at least one of a selection operation, a crossover operation, and a mutation operation.
Information provision device. According to clause 14,
The tree creation unit,
Separating the plurality of questions into a certain number of clusters using attribute information assigned to the plurality of questions,
The attribute information includes information about the correlation between the question and a plurality of data groups.
Information provision device. According to clause 15,
The tree creation unit,
For the cluster currently separated by the genetic algorithm, the final similarity of the currently separated cluster is calculated using the attribute information of the question belonging to the currently separated cluster, and the currently separated cluster is determined based on the final similarity. Determined to be the optimal cluster
Information provision device. delete According to clause 14,
The tree creation unit,
A first question included in a first cluster among the plurality of clusters is generated as a first child node of the parent node, and a second question included in a second cluster different from the first cluster among the plurality of clusters is generated as the parent node. Created as a second child node of a node
Information provision device. According to clause 19,
The parent node is,
Generated from questions included in the first cluster
Information provision device.