Generation method, generation program, and information processing device

Abstract

In this generation method, a computer executes processing for: acquiring a first policy and a second policy which are composed of graph data including a plurality of conditional branches joined by directed edges and nodes each joined to a branch destination of each of the plurality of conditional branches; generating information relating to a first difference between the acquired first and second policies, the first difference being a difference in graph data structure and / or a difference in branching probability of an object allocated to a node of a conditional branch; and generating correspondence information, in which the generated information relating to the first difference is associated with information relating to a second difference between the acquired first and second policies, the second difference being a difference in effect value and / or a difference in degree of certainty.

Description

Generation method, generation program, and information processing device

[0001] The present invention relates to a generation method, a generation program, and an information processing device.

[0002] One type of workflow is the policy flow graph, which diagrams the process of assigning policy target objects, such as people, to services to achieve the policy's objectives in various fields such as medical care, nursing care, and administration.

[0003] For example, various technologies have been proposed to support the formulation of policies, such as predicting the effectiveness of policies. One aspect of policy formulation is that such technologies are used to calculate the effectiveness of multiple policies and select the most effective one from among the multiple policies.

[0004] Japanese Patent Application Laid-Open No. 2021-117837

[0005] However, even if the most effective measure is selected, it is not necessarily easy to change the selected measure. For example, implementing the most effective measure requires resources commensurate with the effect, and there are cases where the current resources are insufficient to meet the required amount, or where it is difficult to increase the resources, or even if it is possible to increase the resources, the cost may be unacceptable.

[0006] In one aspect, the present invention aims to provide a generation method, a generation program, and an information processing device that can realize evaluation of the impact of a policy change.

[0007] In one embodiment of the generation method, a computer executes the following process: acquires a first measure and a second measure composed of graph data including a plurality of conditional branches connected by directed edges and each node connected to the branch destination of each of the plurality of conditional branches; generates information regarding at least one first difference between the difference in the structure of the graph data between the acquired first measure and the second measure or the difference in the branch probability of the object assigned to the node of the conditional branch; and generates correspondence information that associates the generated information regarding the first difference with information regarding at least one second difference between the difference in effect value or the difference in confidence between the acquired first measure and the second measure.

[0008] According to one embodiment, it is possible to evaluate the impact of policy changes.

[0009] FIG. 1 is a block diagram showing an example of the functional configuration of a server device. FIG. 2 is a diagram illustrating an example of a flow graph of a measure. FIG. 3 is a diagram illustrating a specific example of a flow graph of a measure. FIG. 4 is a diagram illustrating one aspect of a problem-solving approach. FIG. 5 is a diagram (1) showing an example of a display of an evaluation value. FIG. 6 is a diagram (2) showing an example of a display of an evaluation value. FIG. 7 is a schematic diagram (1) explaining an example of a numerical calculation of an evaluation value of a region change. FIG. 8 is a schematic diagram (2) explaining an example of a numerical calculation of an evaluation value of a region change. FIG. 9 is a diagram (3) showing an example of a display of an evaluation value. FIG. 10 is a diagram (4) showing an example of a display of an evaluation value. FIG. 11 is a flowchart showing the procedure of the calculation process. FIG. 12 is a diagram showing an example of a hardware configuration.

[0010] Hereinafter, embodiments of the generation method, generation program, and information processing device according to the present application will be described with reference to the accompanying drawings. Each embodiment merely illustrates one example or aspect, and does not limit the range of values, functions, or usage scenarios. Furthermore, each embodiment can be appropriately combined within the scope of not causing any contradiction in the processing content.

[0011] <System Configuration> Fig. 1 is a block diagram showing an example of the functional configuration of a server device 10. The server device 10 shown in Fig. 1 provides a data-based platform that enables sharing, cross-referencing, and updating of policy flow graphs.

[0012] For example, the server device 10 can provide the functions of the data infrastructure platform as a cloud service by executing PaaS (Platform as a Service) type middleware or SaaS (Software as a Service) type applications. The server device 10 corresponds to an example of an information processing device.

[0013] As shown in Fig. 1, the server device 10 can be communicatively connected to the client terminals 30 via a network NW. For example, the network NW may be any type of communication network, whether wired or wireless, such as the Internet or a local area network (LAN). Note that Fig. 1 shows an example in which one client terminal 30 is connected to one server device 10, but any number of client terminals 30 may be connected.

[0014] The client terminal 30 is a terminal device that receives the above-mentioned data infrastructure. For example, the client terminal 30 may be used by a policy planner as an example of an entity that implements a policy, such as a local government or other related party. Note that the client terminal 30 may be realized by any computer, such as a personal computer, a smartphone, a tablet terminal, or a wearable terminal.

[0015] <Flow Graph of Policy> An example of a flow graph of the above policy is shown in Figure 2. Figure 2 is a diagram illustrating an example of a flow graph of a policy. Z1, Z2, Z3, and Z4 in Figure 2 indicate, for example, services that an administrator provides to a user. These may also be referred to as "service implementation components." Specific examples of services include, in the medical field, "interventions" to which the target object of the policy, such as a resident, is assigned, such as undergoing a health checkup or being examined by a specialist, as well as "no intervention" such as follow-up observation, but are not limited to policies in the medical field.

[0016] H1 and H2 indicate, for example, conditional branches including conditions. These may also be referred to as "conditional branch components." Specific examples of conditions, in the medical field, include an estimated glomerular filtration rate (eGFR) below a threshold, a hemoglobin A1c value (HbA1c) below a threshold, and a urinary protein value equal to or greater than a threshold, but are not limited to conditions in the medical field.

[0017] Z1, Z2, Z3, Z4, H1, and H2 may each be referred to as a "component." From the perspective of graph data, such a "component" may correspond to an example of a "node." Hereinafter, a node corresponding to a branch among the nodes may be referred to as a "branch node." Furthermore, a connection between nodes may correspond to an example of an "edge," including a "directed edge."

[0018] In this embodiment, policy planning in the medical field will be described as an example, but the present invention is not limited to this. The above-described embodiment may be used for various policy planning such as work with conditional branching, tests, and questionnaires. In such cases, the same effects as those of the above-described embodiment can be obtained.

[0019] Figure 3 shows a specific example of a policy flow graph. As shown in Figure 3, a policy is modeled as a workflow consisting of a combination of components such as conditional branching and service implementation. Then, the number of people receiving each service is output from a model that has been trained by accumulating information on the flow of people and parameters based on actual values ​​when each conditional branching component is used. This allows local governments to utilize policies implemented by other local governments that are suitable for their own purposes, taking into account their own resources.

[0020] In the example shown in FIG. 3 , the number of people N=1000 is input at S0. At S1, component #1 as service execution component A is set to "health check." At S2, component #2 as conditional branch component B is set to "eGFR<α." If "eGFR<α" is not satisfied (see the NO route at S2), it is determined that "no intervention" by a specialist is required for the citizen, as shown at S5.

[0021] On the other hand, if "eGFR<α" is satisfied (see the YES route at S2), then component #3 as conditional branch component C is set to "HbA1c<β" as shown at S3. If "HbA1c<β" is satisfied (see the YES route at S3), then component #4 as conditional branch component D is set to "nephrologist" as shown at S6, and it is determined that the citizen requires intervention by a "nephrologist." On the other hand, if "HbA1c<β" is not satisfied (see the NO route at S3), then it is determined that the citizen requires intervention by a "diabetes specialist" as shown at S7.

[0022] In the example shown in Figure 3, the number of people who will flow through part #1, part #2, part #3, and part #4 in that order is predicted, as indicated by the arrows. For example, in the policy flow graph shown in Figure 3, the results of assigning the number of people N = 1000 to interventions Z2 to Z4 are as follows: 50 people are assigned to intervention Z2. 150 people are assigned to intervention Z3. Furthermore, 800 people are assigned to intervention Z4.

[0023] 2 and 3, a specific example of another use of the policy flow graph will be described. The server device 10 searches the policy flow graph using attribute information of people in the organization to which the policy planner belongs. The server device 10 identifies the node to which the person is classified among the nodes located at the end of the policy flow graph. This allows the local government that applied the policy to identify medical institutions to recommend to the person, taking into account the health status of the person belonging to the local government and the local government's resources.

[0024] First, the server device 10 identifies people in the organization to which the policy planner belongs. For example, the server device 10 identifies residents of a local government. Next, the server device 10 uses attribute information of people in the organization to which the policy planner belongs to search the output policy flow graph to identify a node to which the identified person is classified among the nodes located at the end of the policy flow graph. The attribute information is biometric information identified by analyzing the person's bodily fluids. The attribute information includes estimated glomerular filtration rate, hemoglobin A1c level, urinary protein level, etc. Note that bodily fluids include blood, lymph, tissue fluid (interstitial fluid, intercellular fluid, interstitial fluid), sweat, tears, nasal mucus, urine, semen, vaginal fluid, amniotic fluid, milk, etc.

[0025] At this time, the server device 10 identifies a node into which the person is classified by comparing the person's attribute information with the conditions included in the conditional branch component. The server device 10 identifies a node into which the identified person is classified among the terminal nodes. The server device 10 then sets the medical institution indicated by the identified node into which the person is classified as a medical institution to be recommended to the identified person. The medical institution indicated by the node is a nephrologist, a diabetes specialist, etc.

[0026] Hereinafter, measures that have already been implemented and have a proven track record will be referred to as "existing measures," while measures that are listed as options for implementation when formulating measures will be referred to as "candidate measures." Furthermore, the flow graph of a measure may be abbreviated as "measure flow." Additionally, among the measure flows, the measure flow corresponding to an existing measure will be referred to as "existing measure flow," and the measure flow corresponding to a candidate measure will be referred to as "candidate measure flow."

[0027] <Data Infrastructure> In the above data infrastructure, policy flows may be shared in any framework. As just one example, the above data infrastructure allows policy flows to be shared among organizations around the world, for example, public organizations such as local governments.

[0028] A policy planner can refer to templates of existing policies from around the world shared in the data infrastructure via the client terminal 30. For example, from the perspective of administrative (political) ease of implementation, a policy candidate can be generated by incorporating portions of flow graphs from multiple existing policies among the templates collected in the data infrastructure. Such flow graphs of policy candidates may be generated automatically or manually using any technology.

[0029] <Automatic Generation of Policy Flow> As just one example, a policy flow can be generated by a data-driven method such as OPT (Optimal Policy Tree) or CF (Causal Forest). With such a data-driven method, a flow graph that increases the effectiveness of a policy is generated preferentially using training data that is assigned a service, for example, an intervention label, to which the correct answer is assigned to the feature of an object used to determine branching conditions.

[0030] As another example, a policy flow can be generated using a library of extracted path structures, which are elements of a flow graph. A "path" here refers to a route formed by a series of branching nodes on a flow graph that leads to a service. For example, a library of path structures can be generated by listing the array of conditions defined for each branching node in a path, i.e., branching sequences, from templates of existing policies collected in the data infrastructure. A policy flow can be generated by searching for a path corresponding to a desired combination of services from this library of path structures.

[0031] <One Aspect of the Problem> As explained in the Background Art section above, evaluation of measures in the above-mentioned conventional techniques tends to place too much emphasis on prediction of effects using SHAP (Shapley Additive exPlanations) values ​​and the like.

[0032] In this way, because the evaluation of measures places too much emphasis on their effectiveness, even if the most effective measure is selected, it is not necessarily easy to change the selected measure. For example, implementing the most effective measure requires resources commensurate with its effectiveness, so there may be cases where the current resources are insufficient to meet the required amount, or where it is difficult to increase the resources, or even if it is possible to increase the resources, the cost may be unacceptable.

[0033] <One aspect of the problem-solving approach> Therefore, in the case of policies that involve people, we focus on the fact that changes affect not only the effectiveness of the policy itself, but also the judgment (consensus-based) of the people who decided on the policy and the structure of its implementation, and affect the scope of resources, such as people and things, that are the target of the policy.

[0034] Based on this perspective, in this embodiment, a calculation function is implemented to calculate evaluation values ​​relating to structural differences in graph data, differences in branching probability, differences in effect values, or differences in the confidence of effect values ​​as indicators for comparing the flow graphs of two measures.

[0035] Figure 4 shows one aspect of the problem-solving approach. As an example, Figure 4 shows an example in which evaluation values ​​for structural change, area change, and effect change are calculated as indicators for comparing an existing policy flow f1 with three candidate policy flows m1 to m3.

[0036] As shown in Figure 4, the evaluation value of a structural change can be calculated based on the cost of converting one of two policy flows into the other. For example, the evaluation value of a structural change can be calculated by calculating the tree edit distance (TED) between the existing policy flow f1 and three policy candidate flows m1 to m3.

[0037] The evaluation value of the area change can be calculated based on the difference in the number of objects allocated by the same branch node between two policy flows. For example, the evaluation value of the area change can be calculated by calculating the increase or decrease in the number of objects input to the same branch node between the existing policy flow f1 and three policy candidate flows m1 to m3.

[0038] The evaluation value of the change in effectiveness can be calculated based on the difference in the degree to which the conditions at the branching nodes affect the effectiveness between two policy flows. For example, the evaluation value of the change in effectiveness can be calculated by calculating the SHAP value of the feature defined as a condition at the branching node between the existing policy flow f1 and three candidate policy flows m1 to m3, that is, the increase or decrease in the contribution to the effectiveness.

[0039] These three types of evaluation values ​​can be displayed, as an example only, as a chart in which the evaluation values ​​for each index are plotted on an axis corresponding to each index, for example, as a radar chart RC1.

[0040] For example, in the radar chart RC1 shown in FIG. 4, in addition to the three indices of structural change, area change, and effect change, the confidence level at the time of calculating the evaluation value of the effect change is plotted as a fourth indices.

[0041] Such a radar chart RC1 allows for multi-perspective evaluation between the existing policy flow f1 and the three policy candidate flows m1 to m3.

[0042] As one aspect, if the evaluation values ​​of the effect change are listed in descending order, they are the evaluation value between the existing policy flow f1 and the candidate policy flow m1, the evaluation value between the existing policy flow f1 and the candidate policy flow m2, and the evaluation value between the existing policy flow f1 and the candidate policy flow m3. From these results, it can be seen that in terms of effectiveness, when changing from the existing policy flow f1 to the candidate policy flow m1, the effect efficiency of the entire policy is most satisfactory.

[0043] From another perspective, if the evaluation values ​​of structural changes are listed in ascending order, they are the evaluation value between the existing policy flow f1 and the candidate policy flow m1, the evaluation value between the existing policy flow f1 and the candidate policy flow m2, and the evaluation value between the existing policy flow f1 and the candidate policy flow m3. From these results, it can be seen that, in terms of empathy, when changing from the existing policy flow f1 to the candidate policy flow m1, the change is most psychologically acceptable.

[0044] As a further aspect, if the evaluation values ​​of the area changes are listed in ascending order, they are the evaluation value between the existing policy flow f1 and the candidate policy flow m1, the evaluation value between the existing policy flow f1 and the candidate policy flow m2, and the evaluation value between the existing policy flow f1 and the candidate policy flow m3. From these results, it can be seen that, in terms of feasibility, changing from the existing policy flow f1 to the candidate policy flow m1 is the easiest to implement.

[0045] These multi-perspective evaluations can prevent the policy candidate flow m1 from being immediately deemed the best just because the change in effect (effectiveness) of the change to the policy candidate flow m1 is the best. In other words, an evaluation can be realized that concludes that the policy candidate flow m1 is the best after confirming that the structural change (empathy) and domain change (feasibility) of the change to the policy candidate flow m1 are also the best. In other words, if the structural change (empathy) and domain change (feasibility) of the change to the policy candidate flow m1 are poorly evaluated, an evaluation can be realized that concludes that the policy candidate flow m1 is not the best.

[0046] Therefore, the calculation function according to this embodiment can evaluate the impact of a policy change. While FIG. 4 illustrates an example in which evaluation values ​​for four indicators are calculated, it is not necessary to calculate evaluation values ​​for all four indicators. For example, it is also possible to calculate an evaluation value for a structural change or an evaluation value for an area change, and an evaluation value for an effect change or an evaluation value for a confidence factor. Furthermore, even when calculating only one of the evaluation values ​​for a structural change or an area change, it is possible to evaluate a policy without placing too much emphasis on the effectiveness of the policy, thereby enabling an evaluation of the impact of a policy change.

[0047] <Configuration of Server Device 10> Fig. 1 shows a schematic diagram of blocks related to the data infrastructure of the server device 10. As shown in Fig. 1, the server device 10 has a communication control unit 11, a storage unit 13, and a control unit 15. Note that Fig. 1 only shows a selection of functional units related to the data infrastructure, and the server device 10 may also be provided with functional units other than those shown.

[0048] The communication control unit 11 is a functional unit that controls communication with other devices such as the client terminal 30. As just one example, the communication control unit 11 can be realized by a network interface card such as a LAN card. In one aspect, the communication control unit 11 receives a policy evaluation request from the client terminal 30 requesting the execution of a policy evaluation, or outputs the results of the policy evaluation to the client terminal 30.

[0049] The storage unit 13 is a functional unit that stores various types of data. As just one example, the storage unit 13 is realized by internal, external, or auxiliary storage of the server device 10. For example, the storage unit 13 stores a policy DB (DataBase) 13A, a feature DB 13B, and an effect DB 13C.

[0050] The policy DB 13A is a database that stores a collection of policy flows. The feature DB 13B is a database that stores feature values ​​related to objects targeted by policies. For example, in the case of a policy in the medical field, in order to assign residents to services, test values ​​for each test item included in health checkup results are stored in the feature DB 13B as feature values ​​used to determine the conditions of branch nodes. Such feature values ​​of residents may be stored by organization, such as by municipality, to which policy planners who have accounts to use the above-mentioned data infrastructure belong. The effect DB 13C is a database that stores the effect of allocating objects to services in the policy flows. For example, for each service, i.e., intervention, the effect of that intervention, such as the degree of improvement in kidney function in the case of a policy flow related to the medical follow-up of CKD (chronic kidney disease), is stored in the effect DB 13C.

[0051] The control unit 15 is a functional unit that performs overall control of the server device 10. For example, the control unit 15 can be realized by a hardware processor. Alternatively, the control unit 15 can be realized by hardwired logic. As shown in FIG. 1 , the control unit 15 includes a receiving unit 15A, an acquiring unit 15B, a calculating unit 15C, and an output unit 15D.

[0052] The reception unit 15A is a processing unit that receives various requests from the client terminal 30. As just one example, the reception unit 15A can receive a policy evaluation request from the client terminal 30, requesting the execution of a policy evaluation. Such a policy evaluation request may include, just one example, a specification of a pair of policy flows to be compared. For example, from among the policy flows stored in the policy DB 13A, an existing policy flow that has already been implemented and has a proven track record and M (any natural number) candidate policy flows can be specified. Furthermore, the policy evaluation request may include a specification of an object that is the target of the policy, or an organization to which the object belongs, such as a local government.

[0053] The acquisition unit 15B is a processing unit that acquires a policy flow from the policy DB 13A. As just one example, the acquisition unit 15B acquires, from the policy DB 13A, the policy flow specified in the policy evaluation request received by the reception unit 15A, i.e., the existing policy flow and M policy candidate flows. Furthermore, the acquisition unit 15B can acquire, from the feature DB 13B, the feature of the object specified in the policy evaluation request. In addition, the acquisition unit 15B can acquire, from the effect DB 13C, the effect when the feature of the object specified in the policy evaluation request is applied to the policy flow specified in the policy evaluation request.

[0054] The calculation unit 15C is a processing unit that calculates an evaluation value for each index. As just one example, the calculation unit 15C calculates evaluation values ​​for N (any natural number) indexes between the existing policy flow and the policy candidate flow for each of M policy candidate flows.

[0055] As one of such indicators, the calculation unit 15C can calculate an evaluation value for a structural change between an existing policy flow and a candidate policy flow. For example, the calculation unit 15C calculates an evaluation value for a structural change based on the cost of converting one policy flow of two policy flows into the other policy flow. As an example, the calculation unit 15C calculates an evaluation value for a structural change between an existing policy flow T １ and policy candidate flow T ２ The tree edit distance Δｓｔｒ (T １ , T ２ ) can be used to calculate the evaluation value of the structural change.

[0056]

[0057] Here, "cost(T １ →T ２ ) is the existing policy flow T １ The node of the policy candidate flow T ２ Furthermore, the cost of replacing the node with the node in the above formula (1) is １ →Λ) is the existing policy flow T １ Furthermore, the cost of deleting a node in the above formula (1) is ２ ) is the policy candidate flow T ２ This refers to the cost of adding a node.

[0058] As another index, the calculation unit 15C can calculate an evaluation value for an area change between an existing policy flow and a policy candidate flow. For example, the calculation unit 15C can calculate an evaluation value for an area change based on the difference in the number of objects allocated by the same branch node between two policy flows. As an example, the calculation unit 15C calculates the area change of the existing policy flow T according to the following formula (2): １ and policy candidate flow T ２ The increase or decrease in the number of objects input to the same branch node between ｃｏｖ (T １ , T ２ ) to calculate the evaluation value of the area change.

[0059]

[0060] Here, "N" in the above formula (2) ２，ｋ (D)" is a flow of policy candidates T ２ In addition, "N" in the above formula (2) refers to the number of people input to branch node k when applying １，ｋ(D)" is the above data set D in the existing policy flow T １ This refers to the number of people input to branch node k when applying

[0061] As a further index, the calculation unit 15C can calculate an evaluation value regarding the change in effect between the existing policy flow and the candidate policy flow. For example, the calculation unit 15C can calculate an evaluation value regarding the change in effect based on the difference in the degree to which the conditions of the branch node affect the effect between the two policy flows. As just one example, the calculation unit 15C calculates the SHAP value of the feature defined as a condition at the branch node between the existing policy flow and the candidate policy flow, i.e., the increase / decrease Δ ｏｕｔ (T １ , T ２ ) to calculate the evaluation value of the effect change.

[0062]

[0063] Here, "φ" in the above formula (3) ｊ (D)" is a flow of policy candidates T ２ In addition, "φ" in the above formula (3) refers to the contribution to the effect of branch node j when applied to ｉ (D)" is the above data set D in the existing policy flow T １ , which refers to the contribution to the effect of branch node i when applied to

[0064] As other indices, the calculation unit 15C can calculate evaluation values ​​of indices relating to the reliability of the existing policy generation model, such as error, variance, precision, loss, and convergence rate. As just one example, the difference in mean regret between the existing policy flow and the candidate policy flow can be calculated as an example of an index of the policy generation model. For example, the mean regret of the existing policy flow can be calculated according to the following formula (4). "max(Y ｔ (l)) is the effect of selecting the most effective measure when applying a possible action flow T to a data set l containing the features of an object group belonging to an arbitrary set, as (Yｔ１ (l)) is the existing policy flow T １ This shows the effect when the formula (4) is applied. In addition, the average regret of the policy candidate flow can be calculated in the same way as the formula (4) below, so the difference between the average regret of the existing policy flow and the average regret of the policy candidate flow can also be calculated.

[0065]

[0066] In addition, the calculation unit 15C can calculate the evaluation value of an index in a digitized measure. For example, an index in a digitized measure refers to an index used when quantitatively comparing formalized rules, documents, etc. For example, an evaluation index used in Rules as Code, legal document comparison, etc. may be used as the index in a digitized measure.

[0067] The output unit 15D is a processing unit that outputs various information to the client terminal 30. As just one example, the output unit 15D outputs to the client terminal 30 the evaluation values ​​of the N indicators calculated by the calculation unit 15C for each of the M policy candidate flows.

[0068] <Specific example (1) of evaluation value output> Fig. 5 is a diagram (1) showing an example of the display of evaluation values. Fig. 5 shows an example in which evaluation values ​​of two indicators, effect change and structural change, between an existing policy flow f11 and two policy candidate flows m11 and m12 are displayed on a client terminal 30. As shown in Fig. 5, the evaluation value of the effect change between the existing policy flow f11 and the policy candidate flow m12 is displayed as "16.8". On the other hand, the evaluation value of the effect change between the existing policy flow f11 and the policy candidate flow m11 is displayed as "16.3".

[0069] Therefore, if we look only at the evaluation value of the effect change, the effect is greatest when changing from the existing policy flow f11 to the policy candidate flow m12, so there is a high possibility that the policy candidate flow m12 will be selected as the policy flow to change from the existing policy flow f11.

[0070] Even if the policy candidate flow m12 is selected in this way, it may be difficult to change the existing policy flow f11 to the policy candidate flow m12. This is because the evaluation value of the structural change between the existing policy flow f11 and the policy candidate flow m12 is large, at "5." For example, when changing from the existing policy flow f11 to the policy candidate flow m12, permission and coordination between related parties are required due to changes in reference data and medical procedures, so the burden of changing to the policy candidate flow m12 is large.

[0071] 5, not only the evaluation value of the effect change but also the evaluation value of the structural change is displayed on the client terminal 30. Therefore, it can be confirmed that the evaluation value of the structural change between the existing policy flow f11 and the policy candidate flow m12 is "5", while the evaluation value of the structural change between the existing policy flow f11 and the policy candidate flow m11 is "1".

[0072] Therefore, the policy planner can confirm that while there is not much difference in the evaluation value of the effect change when changing from the existing policy flow f11 to the candidate policy flow m12 and when changing from the existing policy flow f11 to the candidate policy flow m11, there is a large difference in the evaluation value of the structural change.

[0073] For example, when changing from the existing policy flow f11 to the candidate policy flow m11, the structural change evaluation value is small at "1," so it can be seen that the judgment and execution procedures do not need to be significantly changed from the existing policy. Furthermore, it can be seen that when changing from the existing policy flow f11 to the candidate policy flow m11, it is possible to obtain the same effect as when changing from the existing policy flow f11 to the candidate policy flow m12.

[0074] By making such multi-perspective evaluation possible, it becomes easier to select the candidate policy flow m11 that is the global optimal solution, without falling into a situation where the candidate policy flow m12 that is the local optimal solution is selected as the policy flow to be changed from the existing policy flow f11.

[0075] <Specific Example (2) of Output of Evaluation Values> Fig. 6 is a diagram (2) showing an example of displaying evaluation values. Fig. 6 shows an example in which evaluation values ​​of two indicators, namely, effect change and area change, between an existing policy flow f11 and two policy candidate flows m11 and m12 are displayed on the client terminal 30.

[0076] 6, the evaluation value of the effect change between the existing policy flow f11 and the policy candidate flow m12 is displayed as "16.8", and the evaluation value of the effect change between the existing policy flow f11 and the policy candidate flow m11 is displayed as "16.3". Furthermore, the evaluation value of the area change between the existing policy flow f11 and the policy candidate flow m12 is displayed as "-200", and the evaluation value of the area change between the existing policy flow f11 and the policy candidate flow m11 is displayed as "40".

[0077] In this way, not only the evaluation value of the effect change but also the evaluation value of the area change are displayed on the client terminal 30. Therefore, the policy planner can confirm that while there is not much difference in the evaluation value of the effect change when changing from the existing policy flow f11 to the candidate policy flow m12 and when changing from the existing policy flow f11 to the candidate policy flow m11, there is a large difference in the evaluation value of the area change.

[0078] For example, when changing from the existing policy flow f11 to the candidate policy flow m12, the evaluation value of the area change is large at "-200," so it can be seen that the number of people to be examined or treated, or the amount of change in the number of people to be treated, needs to be significantly changed from the existing policy. On the other hand, when changing from the existing policy flow f11 to the candidate policy flow m11, the evaluation value of the area change is small at "40," so it can be seen that the number of people to be examined or treated, or the amount of change in the number of people to be treated, does not need to be significantly changed from the existing policy. Furthermore, when changing from the existing policy flow f11 to the candidate policy flow m11, it is possible to obtain the same effect as when changing from the existing policy flow f11 to the candidate policy flow m12.

[0079] By making such multi-perspective evaluation possible, it becomes easier to select the candidate policy flow m11 that is the global optimal solution, without falling into a situation where the candidate policy flow m12 that is the local optimal solution is selected as the policy flow to be changed from the existing policy flow f11.

[0080] <Example of Numerical Calculation of Evaluation Value of Area Change> Here, an example of numerical calculation of evaluation value of area change shown in FIG. 6 will be described. FIGS. 7 and 8 are schematic diagrams (1) illustrating an example of numerical calculation of evaluation value of area change. FIG. 7 shows a map M1 indicating the number of people input to each branch node of the existing policy flow f11 and the candidate policy flow m12 for each type of feature defined in the conditions of the branch nodes of the existing policy flow f11 and the candidate policy flow m12. Meanwhile, FIG. 8 shows a map M2 indicating the number of people input to each branch node of the existing policy flow f11 and the candidate policy flow m11 for each type of feature defined in the conditions of the branch nodes of the existing policy flow f11 and the candidate policy flow m11. Note that the map M1 shown in FIG. 7 illustrates an example in which, for a type of feature that exists in a branch node of one policy flow but does not exist in the branch node of the other policy flow, the feature value of the policy flow that does not have the feature is set to zero.

[0081] As shown in Figure 7, the OR set of feature types defined in the branch node conditions between the existing policy flow f11 and the policy candidate flow m12 is: medical examination value A, blood test value B, residence value B, radiology value E, medical history value F, treatment (i), treatment (ii), treatment (iii), surgery (a), and surgery (b). For each feature type, a difference is calculated by subtracting the number of people input into the branch node of the policy candidate flow m12 from the number of people input into the branch node of the existing policy flow f11. The sum of the differences calculated for each feature type is calculated as follows: "500 + 250 - 500 - 300 - 200 + 170 - 170 + 200 - 50 - 100." As a result, the evaluation value of the area change between the existing policy flow f11 and the policy candidate flow m12 is calculated as "-200."

[0082] Fig. 8 is a schematic diagram (2) illustrating an example of numerical calculation of the evaluation value of the area change. Fig. 8 shows a map M2 indicating the number of people input to each branch node of the existing policy flow f11 and the candidate policy flow m11 for each type of feature defined in the conditions of the branch node of each of the existing policy flow f11 and the candidate policy flow m11. Note that map M2 shown in Fig. 8 also shows an example in which, for a type of feature that exists in a branch node of one policy flow but does not exist in the branch node of the other policy flow, the feature value of the policy flow that does not have the feature value is set to zero.

[0083] As shown in Figure 8, the OR set of feature types defined in the branch node conditions between the existing policy flow f11 and the policy candidate flow m11 is: medical examination value A, blood test value B, blood test value C, treatment (i), treatment (ii), and treatment (iii). For each feature type, a difference is calculated by subtracting the number of people input to the branch node of the policy candidate flow m11 from the number of people input to the branch node of the existing policy flow f11. In this way, the sum of the differences calculated for each feature type is calculated as "0 + 250 - 220 + 30 + 0 - 20." As a result, the evaluation value of the area change between the existing policy flow f11 and the policy candidate flow m12 is calculated as "40."

[0084] <Specific Example (3) of Output of Evaluation Values> Fig. 9 is a diagram (3) showing an example of displaying evaluation values. Fig. 9 shows an example in which evaluation values ​​of three indicators, namely, effect change, area change, and structure change, between an existing policy flow f11 and two policy candidate flows m11 and m12 are displayed on the client terminal 30.

[0085] 9, the evaluation values ​​of the three indices, effect change, area change, and structural change, are each normalized to a numerical range from 0 to 1. This normalization makes it possible to compare the evaluation values ​​of each indices with a unified scale.

[0086] Furthermore, the evaluation values ​​for area changes and structural changes are normalized to values ​​in which the magnitude of the evaluation values ​​and the level of the evaluation are positively correlated. In other words, as the change in area (number of people) between the existing policy flow and the candidate policy flow becomes smaller, the evaluation value for area changes increases and the evaluation becomes higher. Also, as the change in structure between the existing policy flow and the candidate policy flow becomes smaller, the evaluation value for structural changes increases and the evaluation becomes higher. This normalization eliminates the need for backward calculations to unify the relationship between indicators in which the magnitude of the evaluation value and the level of the evaluation are positively correlated and indicators in which the magnitude of the evaluation value and the level of the evaluation are negatively correlated, making it possible to more quickly evaluate the impact of policy changes.

[0087] <Specific Example (4) of Output of Evaluation Values> Fig. 10 is a diagram (4) showing an example of displaying evaluation values. Fig. 10 shows an example in which a radar chart RC2 is displayed on the client terminal 30, in which the three types of evaluation values ​​shown in Fig. 9, namely, effect change, area change, and structural change, are plotted on the axes corresponding to each indicator. As shown in Fig. 10, a comprehensive evaluation of the entire multi-perspective indicators can be presented based on the size of the polygon area, allowing the user to intuitively grasp the relative merits of candidate policy flows.

[0088] <Processing Flow> Fig. 11 is a flowchart showing the procedure of the calculation process. This process is merely an example and can be executed when a policy evaluation request for executing a policy evaluation is received from the client terminal 30.

[0089] 11 , when the receiving unit 15A receives a policy evaluation request (step S101), the acquiring unit 15B executes the following process: The acquiring unit 15B acquires the policy flows specified in the policy evaluation request received in step S101, i.e., the existing policy flows and M candidate policy flows, from the policy DB 13A (step S102).

[0090] Next, the calculation unit 15C executes loop processing 1, which repeats the processing of step S103 a number of times corresponding to the number M of policy candidate flows. Furthermore, the calculation unit 15C executes loop processing 2, which repeats the processing of step S103 a number of times corresponding to the number N of types of indicators between the existing policy flow and the m-th policy candidate flow.

[0091] That is, the calculation unit 15C calculates an evaluation value for the nth index between the existing policy flow and the mth policy candidate flow (Step S103).

[0092] By repeating this loop process 2, an evaluation value is calculated for each of the N indicators. Furthermore, by repeating loop process 1, an evaluation value for the N indicators is calculated for each of the M policy candidate flows.

[0093] Thereafter, the output unit 15D outputs the evaluation values ​​of the N indexes calculated for each of the M policy candidate flows in step S103 to the client terminal 30 (step S104), and ends the process.

[0094] <One Aspect of Effect> As described above, the server device 10 according to the present embodiment calculates evaluation values ​​relating to differences in the structure of graph data, differences in branching probability, differences in effect values, or differences in the confidence of effect values ​​as indicators for comparing two flow graphs, one for an existing policy and one for a candidate policy. By outputting evaluation values ​​for multiple indicators among these indicators, it becomes possible to evaluate the flow graphs of the existing policy and the candidate policy from multiple perspectives. Therefore, the server device 10 according to the present embodiment can evaluate the impact of a policy change.

[0095] Although the embodiments of the disclosed device have been described above, the present invention may be embodied in various different forms other than the above-described embodiments. Therefore, other embodiments included in the present invention will be described below.

[0096] <Calculation of Overall Evaluation Value> In the first embodiment described above, an example was given in which an evaluation value for each of a plurality of indicators such as an effect change, an area change, or a structure change is output, but the present invention is not limited to this. For example, the calculation unit 15C can calculate a single overall evaluation value from the evaluation values ​​of a plurality of indicators by performing statistical processing, such as arithmetic averaging or weighted averaging, on the evaluation values ​​of the plurality of indicators.

[0097] <Narrowing Down of Candidate Policies to Be Output> In the above-described first embodiment, all candidate policy flows specified in a policy evaluation request are output. However, the candidate policy flows can be narrowed down to those whose evaluation values ​​for one or more indicators satisfy a specific condition. For example, the output unit 15D can output candidate policy flows whose evaluation value for a specific indicator is greater than or equal to a threshold, or candidate policy flows whose evaluation values ​​for each of multiple indicators are greater than or equal to a threshold. Furthermore, the output unit 15D can output candidate policy flows whose evaluation value for a specific indicator falls within a specified number, for example, the top three, or candidate policy flows whose evaluation values ​​for each of multiple indicators fall within a specified number, for example, the top three.

[0098] <Distribution and Integration> Furthermore, the components of each device shown in the figure do not necessarily have to be physically configured as shown. In other words, the specific form of distribution and integration of each device is not limited to that shown in the figure, and all or part of the devices can be functionally or physically distributed and integrated in any unit depending on various loads, usage conditions, etc. For example, the reception unit 15A, acquisition unit 15B, calculation unit 15C, or output unit 15D may be connected via a network as an external device to the server device 10. Furthermore, the reception unit 15A, acquisition unit 15B, calculation unit 15C, or output unit 15D may each be included in a separate device, and the functions of the server device 10 may be realized by the devices being connected to a network and operating together.

[0099] <Hardware Configuration> The various processes described in the above embodiments can be realized by executing a prepared program on a computer such as a personal computer, a workstation, etc. Therefore, an example of a computer that executes a calculation program having the same functions as those in the first and second embodiments will be described below with reference to FIG.

[0100] Fig. 12 is a diagram showing an example of a hardware configuration. As shown in Fig. 12, a computer 100 has an operation unit 110a, a speaker 110b, a camera 110c, a display 120, and a communication unit 130. The computer 100 also has a CPU 150, a ROM 160, a HDD 170, and a RAM 180. These units 110 to 180 are connected via a bus 140.

[0101] 12, a calculation program 170a that performs the same functions as the reception unit 15A, acquisition unit 15B, calculation unit 15C, and output unit 15D shown in the first embodiment is stored in the HDD 170. This calculation program 170a may be integrated or separated, similar to the components of the reception unit 15A, acquisition unit 15B, calculation unit 15C, and output unit 15D shown in FIG. 1. In other words, it is not necessary for all of the data shown in the first embodiment to be stored in the HDD 170, as long as the data used for processing is stored in the HDD 170.

[0102] Under such an environment, the CPU 150 reads the calculation program 170a from the HDD 170 and loads it into the RAM 180. As a result, the calculation program 170a functions as a calculation process 180a, as shown in FIG. 12 . The calculation process 180a loads various data read from the HDD 170 into an area of ​​the storage area of ​​the RAM 180 allocated to the calculation process 180a, and executes various processes using the loaded data. For example, examples of the processes executed by the calculation process 180a include the processes shown in FIG. 11 . Note that the CPU 150 does not necessarily need to operate all of the processing units shown in the first embodiment; it is sufficient that the processing units corresponding to the processes to be executed are virtually implemented.

[0103] The calculation program 170a does not necessarily have to be stored in the HDD 170 or the ROM 160 from the beginning. For example, each program may be stored on a "portable physical medium" such as a flexible disk, a so-called FD, a CD-ROM, a DVD disk, a magneto-optical disk, or an IC card that is inserted into the computer 100. The computer 100 may then acquire and execute each program from such a portable physical medium. Alternatively, each program may be stored in another computer or server device connected to the computer 100 via a public line, the Internet, a LAN, a WAN, or the like, and the computer 100 may acquire and execute each program from such a computer or server device.

[0104] REFERENCE SIGNS LIST 10 Server device 11 Communication control unit 13 Storage unit 13A Policy DB 13B Feature DB 13C Effect DB 15 Control unit 15A Reception unit 15B Acquisition unit 15C Calculation unit 15D Output unit 30 Client terminal

Claims

1. Acquire a first measure and a second measure configured from graph data including a plurality of conditional branches connected by directed edges and nodes connected to branch destinations of the plurality of conditional branches; generating information about at least one of a difference in the structure of graph data between the acquired first measure and the acquired second measure and a difference in branch probability of an object assigned to a node of the conditional branch; generating correspondence information in which information on a second difference, which is at least one of a difference in effect value or a difference in confidence between the acquired first measure and the acquired second measure, is associated with the generated information on the first difference; A generating method characterized in that the processing is carried out by a computer.

2. The generation method described in claim 1, characterized in that the process of calculating the first difference includes a process of calculating the difference in structure of the graph data based on the cost of converting one of the graph data of the first measure and the graph data of the second measure into the other measure.

3. The method of claim 2 , wherein the cost corresponds to a tree edit distance.

4. The generation method according to claim 1, characterized in that the process of calculating the first difference includes a process of calculating the difference in branch probability based on the difference in the number of objects assigned by the same conditional branch node between the first measure and the second measure.

5. The generation method described in claim 1, characterized in that the process of calculating the second difference includes a process of calculating the difference in the effect value based on the difference in the degree of influence that the multiple conditional branches have on the effect value between the first measure and the second measure.

6. The generation method according to claim 1, characterized in that the computer further executes a process of outputting a radar chart in which the differences in the structure of the graph data, the differences in the branching probabilities, and at least one of the differences in the effect values ​​and the differences in the confidence factors are plotted in a polygonal shape.

7. Acquire a first measure and a second measure configured from graph data including a plurality of conditional branches connected by directed edges and nodes connected to branch destinations of the plurality of conditional branches; generating information about at least one of a difference in the structure of graph data between the acquired first measure and the acquired second measure and a difference in branch probability of an object assigned to a node of the conditional branch; generating correspondence information in which information on a second difference, which is at least one of a difference in effect value or a difference in confidence between the acquired first measure and the acquired second measure, is associated with the generated information on the first difference; A generating program that causes a computer to execute a process.

8. Acquire a first measure and a second measure configured from graph data including a plurality of conditional branches connected by directed edges and nodes connected to branch destinations of the plurality of conditional branches; generating information about at least one of a difference in the structure of graph data between the acquired first measure and the acquired second measure and a difference in branch probability of an object assigned to a node of the conditional branch; generating correspondence information in which information on a second difference, which is at least one of a difference in effect value or a difference in confidence between the acquired first measure and the acquired second measure, is associated with the generated information on the first difference; An information processing device comprising a control unit that executes processing.