WO2022215435A1 - 行動変容促進装置 - Google Patents
行動変容促進装置 Download PDFInfo
- Publication number
- WO2022215435A1 WO2022215435A1 PCT/JP2022/011025 JP2022011025W WO2022215435A1 WO 2022215435 A1 WO2022215435 A1 WO 2022215435A1 JP 2022011025 W JP2022011025 W JP 2022011025W WO 2022215435 A1 WO2022215435 A1 WO 2022215435A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- risk
- information
- user
- unit
- causal
- Prior art date
Links
- 230000008859 change Effects 0.000 title claims abstract description 37
- 230000003542 behavioural effect Effects 0.000 title claims abstract description 12
- 230000001364 causal effect Effects 0.000 claims abstract description 150
- 230000006399 behavior Effects 0.000 claims description 53
- 230000004048 modification Effects 0.000 claims description 26
- 238000012986 modification Methods 0.000 claims description 26
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000001737 promoting effect Effects 0.000 claims description 6
- 230000009471 action Effects 0.000 claims description 3
- 238000000034 method Methods 0.000 description 35
- 230000008569 process Effects 0.000 description 24
- 238000004891 communication Methods 0.000 description 13
- 238000000605 extraction Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 239000000284 extract Substances 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000006249 magnetic particle Substances 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
Definitions
- One aspect of the present invention relates to a behavior change promotion device.
- One aspect of the present invention has been made in view of the above circumstances, and aims to encourage users to effectively change their behavior.
- a behavior modification promotion device is a behavior modification promotion device that presents information for reducing the risk of an unforeseen situation to a user, comprising learning user information about the user, user A first learning unit that builds a first learning model for estimating the user's risk by learning in association with risk information that is information related to the risk of Build a second learning model for estimating the risk causal effect, which is the degree of increase in risk associated with having risk factors, by learning information related to risk factors that affect risk and risk information in association with each other. a second learning unit for estimating the user's risk by inputting estimation user information related to the user into the first learning model; and estimating the risk causal effect using the second learning model. a second estimation unit, an advice generation unit that generates advice information including at least the risk estimated by the first estimation unit and the risk causal effect estimated by the second estimation unit, and an output unit that outputs the advice information , provided.
- a first learning model for estimating risk from user information and risk information is constructed, and risk factors are calculated from information related to risk factors contained in user information and risk information.
- a second learning model is constructed that estimates risk causality, which is the degree of increased risk associated with having. Then, in the behavior modification promoting device according to one aspect of the present invention, the user's risk is estimated by inputting the user information for estimation into the first learning model, and the risk causal effect is calculated using the second learning model. Advice information including the estimated risk and risk causal effect is generated and the advice information is output.
- a second learning model for estimating the risk causal effect which is the degree of increase in risk associated with having a risk factor, is constructed. and the second learning model estimates a risk causal effect according to a given risk factor.
- FIG. 4 is a diagram showing an example of advice master information and advice information; It is a figure explaining step-by-step output of advice information. It is a flowchart which shows the learning process which the behavior change promotion apparatus which concerns on this embodiment implements. It is a flowchart which shows the estimation process which the behavioral change promotion apparatus which concerns on this embodiment implements. It is a figure which shows the hardware constitutions of the behavior change promotion apparatus which concerns on this embodiment.
- FIG. 1 is a functional block diagram of a behavior change promotion device 1 according to this embodiment.
- the behavior change promotion device 1 is a device that presents information for reducing the risk of unforeseen circumstances to a predetermined user.
- the behavior change promotion device 1 according to the present embodiment presents information for reducing the risk of an accident during driving, which is an unforeseen situation that can occur to a user (that is, a driver) who drives.
- the risk of an accident is, for example, whether or not an accident has occurred, the number of times an accident has occurred, or the magnitude of damage when an accident occurs. In the following, an example in which the "risk of accident" is "presence or absence of occurrence of an accident" will be described.
- the behavior modification promotion device 1 is provided, for example, to communicate with a communication device (controller) of a vehicle driven by each user, and transmits information for reducing risk to the communication device of the vehicle driven by each user ( present).
- the behavior modification promotion device 1 may transmit (present) information for reducing risk by e-mail or the like to a terminal such as a smartphone held by each user, for example.
- the behavior modification promotion device 1 includes a user information DB 11, a feature amount extraction unit 12, a risk information DB 13, a learning unit 14 (first learning unit, second learning unit), and a model DB 15. , an estimation unit 16 (a first estimation unit, a second estimation unit, a calculation unit), an estimation information DB 17, an advice master information DB 18, an advice generation unit 19, an advice information DB 20, and an output unit 21. ing.
- the user information DB 11 is a database that stores user information, which is information related to each user.
- User information largely includes mobility data and non-mobility data.
- Mobility data is data about driving of a user.
- Non-mobility data is user data that is not directly related to driving, such as user attribute information and behavior information.
- Mobility data is collected, for example, from sensors attached to the vehicle, drive recorders, questionnaire information, or the like.
- the non-mobility data is collected, for example, from information equipment owned by the user, service usage logs, questionnaire information, or the like.
- FIG. 4A is an example of a part of user information stored in the user information DB 11. FIG. In the example shown in FIG.
- a personal identifier that uniquely identifies the user, driving time in snow, highway driving time, age, and region are associated with each other and stored. It is In such user information, age and region are non-mobility data, and snow-covered driving time and expressway driving time are mobility data.
- age and region are non-mobility data
- snow-covered driving time and expressway driving time are mobility data.
- A is an area (for example, a residential area).
- the feature amount extraction unit 12 is a function that extracts feature amounts from the user information stored in the user information DB 11 .
- the feature amount extraction unit 12 extracts feature amounts from the user information based on any predetermined rule.
- the feature amount extraction unit 12 may use part or all of the user information stored in the user information DB 11 as the feature amount, or may apply predetermined processing to part or all of the user information stored in the user information DB 11. You may extract the feature-value produced
- the feature quantity extraction unit 12 uses at least the person identifier as a key and extracts the associated feature quantity (record for each person identifier).
- the feature amount extraction unit 12 may use the person identifier and the date/time identifier as keys to extract the associated feature amount (a record for each combination of the person identifier and the date/time identifier).
- the timing of the feature amount extraction processing by the feature amount extraction unit 12 may be arbitrary timing, and may be repeatedly performed at predetermined time intervals.
- the risk information DB 13 is a database that stores risk information indicating whether there is a risk (whether an accident has occurred) for each user.
- FIG. 4(b) is an example of part of the risk information stored in the risk information DB 13. As shown in FIG. In the example shown in FIG. 4B, as risk information, a person identifier that uniquely identifies a user and the presence or absence of risk are associated with each other and stored.
- the learning unit 14 performs first learning for constructing a prediction model (first learning model) for estimating the user's risk, and for estimating the risk causal effect, which is the degree of increase in risk associated with having a risk factor. Second learning for constructing a causal model (second learning model).
- the timing at which the prediction model and the causal model are constructed by the learning unit 14 may be any timing.
- the learning unit 14 may store the new prediction model or causal model in the model DB 15 so as to replace the existing prediction model or causal model.
- the model DB 15 stores prediction models and causal models generated by the learning unit 14 .
- the learning unit 14 learns by associating user information (learning user information) and risk information with each other, and builds a prediction model for estimating the user's risk.
- the user information here is, in detail, the feature amount extracted by the feature amount extraction unit 12 .
- Risk information is risk information stored in the risk information DB 13 .
- the learning unit 14 generates learning data by associating the feature amount and the risk information with, for example, a person identifier as a key.
- the date and time identifier of the feature amount and the date and time identifier of the risk information may be different from each other.
- the feature amount may be data observed in a period of N years
- the linked risk information may be data observed in a period of N+1 years.
- FIG. 2 is a diagram showing a first learning processing image (upper stage) and an estimation processing image (lower stage) related to risk estimation.
- first learning feature quantities in which snowy driving time, highway driving time, age, and region are associated with each other, and risk information, which is an objective variable, are linked.
- risk information which is an objective variable
- learning is carried out and a prediction model is constructed.
- the construction of the prediction model may be carried out, for example, using existing supervised learning algorithms, for example, existing statistical methods and machine learning methods for predicting risk from feature values (logistic regression, gradient boosting decision tree, neural networks, etc.).
- a prediction model is constructed using a classification model, and the risk information indicates the magnitude of damage when an accident occurs.
- a prediction model may be constructed using a regression model when indicated by numerical data such as
- the learning unit 14 learns information related to risk factors that are included in the user information (learning user information) and that affect the risk, and the risk information in association with each other, and learns the risk factors.
- Build a causal model to estimate the risk causal effect which is the degree of increased risk associated with having Information related to risk factors includes, for example, feature amounts that are risk factors and feature amounts that affect at least one of the risk factors and risks.
- the learning unit 14 associates and learns a feature amount that is a risk factor, a feature amount that affects at least one of the risk factor and the risk, and risk information, and builds a causal model.
- the information related to the risk factor here (that is, the feature amount that is the risk factor, and the feature amount that affects at least one of the risk factor and the risk) is, in detail, the feature amount extracted by the feature amount extraction unit 12. quantity.
- the risk information is risk information stored in the risk information DB 13 .
- the learning unit 14 generates learning data by associating a feature amount that is a risk factor, a feature amount that affects at least one of the risk factor and the risk, and risk information with, for example, a person identifier as a key. When linking, the date and time identifier of each feature quantity and the date and time identifier of risk information may be different from each other.
- each feature amount may be data observed in a period of N years, and the associated risk information may be data observed in a period of N+1 years.
- the feature amount that affects at least one of the risk factor and the risk may not be used. That is, the learning unit 14 may associate the feature quantity, which is the risk factor, with the risk information, learn them, and build a causal model.
- FIG. 3 is a diagram showing a second learning processing image (upper stage) and an estimation processing image (lower stage) regarding risk causal effect estimation.
- the second learning the high speed Road driving time, age, and region are selected, linked with risk information, learning is performed, and a causal model is constructed.
- Such a causal model calculates the risk causal effect for each user in a predetermined risk factor based on the feature quantity (the risk factor or the feature quantity that affects at least one of the risks) extracted from the information included in the user information. is a model.
- the causal model makes it possible to estimate the degree of increase in risk (risk causal effect) due to possession of a predetermined risk factor in a predetermined user, for example, from the risk factor or the feature value that affects at least one of the risks.
- existing causal inference methods such as Meta-Learners.
- other causal inference methods such as regression models, the average risk causal effect of users may be calculated instead of calculating the risk causal effect for each user.
- a feature quantity for a given risk factor (a feature quantity that affects at least one of the risk factor and the risk) may be selected by any method.
- Causal models may be constructed for the number of types of risk factors.
- the risk factors are, for example, "60 minutes or more driving time in snow", “60 minutes or more driving time on expressway", and the like.
- a single risk factor may be formed by combining a plurality of risk factors. That is, for example, "60 minutes or more driving time in snow and 60 minutes or more driving time on expressway" may be regarded as one risk factor.
- the estimating unit 16 is a function that performs first estimation for estimating the user's risk using the prediction model and second estimation for estimating the risk causal effect using the causal model.
- the timing of risk estimation and risk causal effect estimation by the estimation unit 16 may be any timing.
- the estimation unit 16 generates estimated risk information derived (estimated) by the first estimation (see FIG. 5A) and causal effect information derived (estimated) by the second estimation (information including risk causal effect, etc.) (see FIG. 5(b)) is stored in the estimated information DB 17.
- the estimated information DB 17 stores estimated risk information and causal effect information estimated by the estimation unit 16 .
- the estimation unit 16 estimates the user's risk indicated by the user information by inputting the user information (estimation user information) into the prediction model stored in the model DB 15 .
- the user information here is, in detail, the feature amount extracted by the feature amount extraction unit 12 .
- Estimated risk information indicating FIG. 5A is a diagram showing an example of estimated risk information.
- estimated risk information is information in which a person identifier that uniquely identifies a user and a value indicating risk (probability, score, etc.) are associated with each other.
- the estimation unit 16 estimates the risk causal effect using the causal model. Specifically, the estimating unit 16 determines that at least one of the risk factor and the risk that is included in the information that is included in the user information (estimating user information) and is related to the risk factor that affects the risk. By inputting the feature quantity into the causal model stored in the model DB 15, the risk causal effect of a given risk factor for a given user is estimated. In addition, when calculating the average risk causal effect of users instead of calculating the risk causal effect for each user using other causal inference methods such as regression models, parameters of the trained causal model (risk The user's average risk causal effect may be calculated by referring to the factor regression coefficient).
- the estimating unit 16 refers to the information on whether or not a predetermined risk factor is possessed from the user information of the predetermined user, and compares the output result of the causal model for the risk factor and the information on whether or not the risk factor is possessed. Based on this, the risk causal effect in the user may be derived for the risk factors held by the user.
- the feature value that affects at least one of the risk factors or risks, in which the highway driving time, age, and region are associated with each other is input to the causal model.
- the risk causal effect which is the degree of increase in risk associated with having a given risk factor.
- Information such as age and region is a feature quantity that affects at least one of risk factors and risks, and is information that indicates the user's features.
- the estimating unit 16 may further calculate the degree of influence of the feature quantity that affects at least one of the risk factor and the risk on the risk causal effect. That is, in the second estimation, the estimating unit 16 calculates the degree of influence of the user's features, which are the feature amounts that affect at least one of the risk factor and the risk, on the risk causal effect, and uses the degree of influence as the causal effect information. (See FIG. 5(b)).
- a user's feature which is a feature quantity that affects at least one of risk factors and risks, is information such as the age and region of the user included in the user information, for example.
- an algorithm linear model, gradient boosting decision tree, etc. that can calculate the importance of feature values used for learning a model that estimates the risk causal effect for each user. , it is possible to calculate how much a predetermined feature amount (a feature amount that affects at least one of risk factors and risks) influences the increase in risk causal effect from the trained model.
- FIG. 5B is a diagram showing an example of causal effect information including risk causal effects derived by the estimation unit 16. As shown in FIG. In the example shown in FIG. 5(b), the causal effect information includes, in addition to the risk causal effect, age and region indicating user characteristics, which are feature quantities that affect at least one of risk factors and risks. Contains the degree of influence on the effect.
- the risk factors include "60 minutes or more driving time during snow cover” and "60 minutes or more driving time on expressway”.
- risk causal Effects and the impact of each feature (age and region) on risk causality are derived.
- the advice master information DB 18 is a database that stores advice master information in which risk factors and advice content are associated.
- FIG. 6A is a diagram showing an example of advice master information.
- the advice content of "Let's keep more distance between vehicles than usual” is associated with the risk factor of "Driving time of 60 minutes or more in snowy conditions”. Further, the content of the advice "Let's use the navigation" is associated with the risk factor of "60 minutes or more driving time on the highway”.
- the advice generating unit 19 generates advice information including at least the risk estimated by the estimating unit 16 in the first estimation and the risk causal effect estimated by the estimating unit 16 in the second estimation.
- the advice generation unit 19 associates the risk and the risk causal effect with at least the person identifier as a key, and acquires them from the estimated information DB 17 . Further, in detail, the advice generation unit 19 calculates the degree of influence of the user's features (feature values affecting at least one of the risk factor and the risk) on the risk causal effect estimated by the estimation unit 16 in the second estimation. It also generates advisory information to include.
- the advice generator 19 acquires the degree of influence of the user's characteristics on the risk causal effect from the estimated information DB 17 .
- the advice generating unit 19 is information associated in advance with the risk factor related to the risk causal effect estimated in the second estimation, and indicates an action to prompt the user to reduce the risk. generates advice information that further includes behavior change promotion content.
- the advice generation unit 19 acquires advice content (behavior change promotion content) associated with the risk factor by referring to the advice master information DB 18 .
- the timing of advice information generation by the advice generation unit 19 may be any timing.
- the advice generation unit 19 stores the generated advice information in the advice information DB 20 .
- the advice information DB 20 stores advice information.
- FIG. 6(b) is a diagram showing an example of advice information.
- the estimated risk of the user indicated by the person identifier "XXXX” is "80”
- the risk factor is "60 minutes or more of driving time in snow”
- the risk causal effect is " 30”
- region is A” for the risk factor is “age is 10”
- the advice content corresponding to the risk factor is “more distance between vehicles than usual”. Let's keep our distance.”
- the output unit 21 outputs advice information stored in the advice information DB 20 .
- the output unit 21 transmits, for example, advice information to the communication device (controller) of the vehicle driven by the user indicated by the person identifier.
- the advice information is then presented on the display of the vehicle driven by the user.
- the advice information may be output by voice in the vehicle driven by the user.
- the output unit 21 may transmit, for example, the advice information to an information device such as a smart phone held by the user indicated by the person identifier by e-mail or the like.
- the output unit 21 may output the advice information step by step. It is also possible to output the risk causal effect obtained by outputting, and finally output the advice content (behavioral change promotion content). In addition, regarding the advice information, after outputting the risk causal effect in the above and before outputting the advice content (behavior change promotion content), the output unit 21 outputs the user's characteristics for the risk causal effect estimated by the estimating unit 16. (feature quantity affecting at least one of risk factor and risk) may be output.
- FIG. 7 is a diagram explaining stepwise output of advice information.
- the message "Your accident risk score is [8]" is first output as the information indicating the estimated risk.
- information indicating the risk causal effect “One of the factors that increase your risk is [driving on unfamiliar roads]. is believed to have increased by [2].” is output.
- risk causal effect “One of the factors that increase your risk is [driving on unfamiliar roads]. is believed to have increased by [2].” is output.
- only the risk causal effect for a single risk factor is shown, but only risk factors with high risk causal effects may be shown, or risk causal effects for multiple risk factors may be shown. may be shown.
- information indicating the degree of influence of the user's characteristics on the estimated risk causal effect "People who are likely to increase their risk especially by [driving on unfamiliar roads] are considered to have the following characteristics.
- FIG. 8 the learning processing and estimation processing performed by the behavior change promotion device 1 according to this embodiment will be described with reference to FIGS. 8 and 9.
- FIG. 8 is a flowchart showing the learning process performed by the behavior change promotion device 1.
- FIG. 8 Note that the processes of steps S2 and S3 and the processes of steps S4 and S5 shown in the flowchart of FIG. 8 do not necessarily have to be executed in the order shown in FIG. That is, the processes of steps S4 and S5 may be executed before the processes of steps S2 and S3, or may be executed simultaneously with the processes of steps S2 and S3. Similarly, the processes of steps S12 and S13 and the processes of steps S14 and S15 shown in the flowchart of FIG. 9 do not necessarily have to be executed in the order shown in FIG.
- steps S14 and S15 may be executed before the processes of steps S12 and S13, or may be executed simultaneously with the processes of steps S12 and S13.
- the learning process first, feature amounts are extracted from the user information for learning (step S1). Subsequently, the feature amount and the risk information are learned in association with each other, and a prediction model for estimating the user's risk is constructed (step S2). The prediction model is stored in the model DB 15 (step S3).
- the feature amount that is the information included in the learning user information and is a risk factor that affects the risk is learned in association with each other.
- a causal model is constructed for estimating the risk causal effect, which is the degree of increase in risk associated with having a predetermined risk factor (step S4).
- a causal model for estimating the risk causal effect may be constructed by learning the feature values of the risk factors and the risk information in association with each other.
- the causal model is stored in the model DB 15 (step S5). The above is the learning process.
- FIG. 9 is a flowchart showing the estimation process performed by the behavior change promotion device 1.
- FIG. 9 in the estimation process, first, feature amounts are extracted from the user information for estimation (step S11). Subsequently, the user's risk is estimated by inputting the feature quantity into the prediction model (step S12). The estimated risk is stored in the estimation information DB 17 (step S13).
- the risk is determined by inputting into the causal model the feature amount that affects at least one of the risk factor and the risk, which is included in the user information for estimation and is information related to the risk factor that affects the risk.
- the causal effect is estimated, and the degree of influence of the user's feature (feature value affecting at least one of the risk factor and the risk) on the risk causal effect is calculated from the learned causal model parameters (step S14).
- feature value affecting at least one of the risk factor and the risk is calculated from the learned causal model parameters.
- risk causal effects may be estimated.
- the estimated risk causal effects and the like are stored in the estimated information DB 17 (step S15).
- advice information is generated based on the information in the estimated information DB 17 and the information in the advice master information DB 18 (step S16).
- the generated advice information is stored in the advice information DB 20 (step S17).
- the advice information is presented (output) to the user (step S18). The above is the estimation processing.
- the behavior modification promotion device 1 is a behavior modification promotion device that presents information for reducing the risk of unforeseen situations to the user, and is a user information for learning about the user, and the user's Risk information, which is information related to risk, is associated with each other and learned to build a predictive model for estimating the user's risk, and the information included in the learning user information that is the risk factor that affects the risk
- the learning unit 14 for building a causal model for estimating the risk causal effect which is the degree of increase in risk associated with having a risk factor, by learning information related to and risk information in association with each other, and a learning unit 14 related to the user
- An estimating unit 16 that estimates a user's risk by inputting user information for estimation into a prediction model and estimates a risk causal effect using a causal model, the risk estimated by the estimating unit 16, and an estimating unit 16 for generating advice information including at least the risk causal effect estimated by 16, and an output unit 21 for outputting the advice information.
- a prediction model for estimating risk from user information and risk information is constructed, and information on risk factors included in user information and risk factors are determined from risk information.
- a causal model is constructed that estimates the risk causal effect, which is the degree of increase in risk involved.
- the user's risk is estimated by inputting the estimation user information into the prediction model, and the risk causal effect is estimated by using the causal model, Advice information including risks and risk causal effects is generated and the advice information is output.
- a causal model for estimating the risk causal effect which is the degree of increase in risk associated with having a risk factor, is constructed.
- Causal models estimate risk causal effects according to given risk factors. By generating and outputting advice information including such risk causal effects, the user is presented with how much the risk increases due to having the risk factor. It is possible to accurately grasp the impact on As a result, it is possible to encourage the user to effectively change behavior against risk. In addition, by centrally estimating risks and risk causal effects, it is possible to improve processing efficiency related to promotion of behavioral change.
- the information related to the risk factor includes a feature amount that is the risk factor and a feature amount that affects at least one of the risk factor and the risk, and the learning unit 14 learns the feature amount that is the risk factor and the risk factor or the risk.
- a causal model may be constructed by learning by associating the feature amount that affects at least one side with the risk information. This makes it possible to build a causal model that can more appropriately estimate the risk causal effect.
- the feature quantity that affects at least one of the risk factor and the risk may include information that indicates the features of the user. Thereby, it is possible to estimate the risk causal effect considering the characteristics of the user.
- the estimating unit 16 inputs into the causal model the risk factor or the feature amount that affects at least one of the risk, which is included in the information included in the user information for estimation and is related to the risk factor that affects the risk. may estimate the risk causal effect. By inputting information related to risk factors into the causal model in this way, the risk causal effect can be estimated more appropriately.
- the estimating unit 16 may calculate the degree of influence of the feature quantity that affects at least one of the risk factor and the risk on the risk causal effect. In addition, the estimating unit 16 calculates the degree of influence of the feature amount that influences at least one of the risk factor and the risk calculated based on the causal model on the risk causal effect, and the degree of influence on at least one of the risk factor and the risk in the user. Based on the information on whether or not the feature value is provided, the degree of influence of the feature value possessed by the user (feature value that affects at least one of the risk factor and the risk) on the risk causal effect may be calculated. .
- the degree of influence of the feature amount (feature amount that affects at least one of the risk factor and the risk) with respect to the risk causal effect, for example, the user's characteristics that the risk causal effect tends to increase can be calculated. becomes possible.
- the advice generation unit 19 may generate advice information that further includes the degree of influence of the user's characteristics, which is a feature amount that affects at least one of the risk factor and the risk for the risk causal effect calculated by the estimation unit 16. .
- the causal model calculates the degree of influence of the user's characteristics on the risk causal effect, and generates and outputs advice information including the degree of influence of the user's characteristics. and the degree of influence thereof will be presented to the user. Then, based on the degree of influence of the predetermined feature amount on the risk causal effect of the predetermined risk factor calculated above and information on whether or not the predetermined user possesses the feature amount, the You may calculate the influence of the feature-value which a user has.
- the user By presenting such information to the user, the user is made to understand how his or her characteristics affect the risk causal effect, and to understand that the increase in risk is an event unique to his or her characteristics. be able to. This makes it possible to more appropriately prompt the user for effective behavioral change against risk.
- the advice generating unit 19 generates behavior change promotion content, which is information associated in advance with the risk factor related to the risk causal effect estimated by the estimating unit 16 and indicates an action to prompt the user to reduce the risk.
- Advisory information may be generated that further includes.
- the output unit 21 first outputs the risk estimated by the estimation unit 16, then outputs the risk causal effect estimated by the estimation unit 16, and then outputs the risk causal effect estimated by the estimation unit 16.
- the degree of influence of the user's characteristics on the effect may be output, and finally the behavior modification promotion content may be output.
- the output unit 21 does not have to output the degree of influence of the user's characteristics.
- the behavior modification promotion device 1 may be physically configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, a bus 1007 and the like.
- the term “apparatus” can be read as a circuit, device, unit, or the like.
- the hardware configuration of the behavior change promotion device 1 may be configured to include one or more of each device shown in the figure, or may be configured without including some of the devices.
- Each function in the behavior modification promotion device 1 is performed by loading predetermined software (program) on hardware such as the processor 1001 and the memory 1002, the processor 1001 performs calculation, communication by the communication device 1004, memory 1002 and It is realized by controlling reading and/or writing of data in the storage 1003 .
- the processor 1001 for example, operates an operating system and controls the entire computer.
- the processor 1001 may be configured with a central processing unit (CPU) including an interface with peripheral devices, a control device, an arithmetic device, registers, and the like.
- CPU central processing unit
- the control functions of the learning unit 14 and the like may be implemented by the processor 1001 .
- the processor 1001 also reads programs (program codes), software modules and data from the storage 1003 and/or the communication device 1004 to the memory 1002, and executes various processes according to them.
- programs program codes
- software modules software modules
- data data from the storage 1003 and/or the communication device 1004
- the program a program that causes a computer to execute at least part of the operations described in the above embodiments is used.
- control functions of the learning unit 14 and the like may be implemented by a control program stored in the memory 1002 and operated by the processor 1001, and other functional blocks may be similarly implemented. Although it has been described that the above-described various processes are executed by one processor 1001, they may be executed by two or more processors 1001 simultaneously or sequentially. Processor 1001 may be implemented with one or more chips. Note that the program may be transmitted from a network via an electric communication line.
- the memory 1002 is a computer-readable recording medium, and is composed of at least one of, for example, ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), RAM (Random Access Memory), etc. may be
- ROM Read Only Memory
- EPROM Erasable Programmable ROM
- EEPROM Electrical Erasable Programmable ROM
- RAM Random Access Memory
- the memory 1002 may also be called a register, cache, main memory (main storage device), or the like.
- the memory 1002 can store executable programs (program codes), software modules, etc. for implementing a wireless communication method according to an embodiment of the present invention.
- the storage 1003 is a computer-readable recording medium, for example, an optical disc such as a CDROM (Compact Disc ROM), a hard disk drive, a flexible disc, a magneto-optical disc (for example, a compact disc, a digital versatile disc, a Blu-ray (registered disk), smart card, flash memory (eg, card, stick, key drive), floppy disk, magnetic strip, and/or the like.
- Storage 1003 may also be called an auxiliary storage device.
- the storage medium described above may be, for example, a database, server, or other suitable medium including memory 1002 and/or storage 1003 .
- the communication device 1004 is hardware (transmitting/receiving device) for communicating between computers via a wired and/or wireless network, and is also called a network device, network controller, network card, communication module, etc., for example.
- the input device 1005 is an input device (for example, keyboard, mouse, microphone, switch, button, sensor, etc.) that receives input from the outside.
- the output device 1006 is an output device (eg, display, speaker, LED lamp, etc.) that outputs to the outside. Note that the input device 1005 and the output device 1006 may be integrated (for example, a touch panel).
- Each device such as the processor 1001 and the memory 1002 is connected by a bus 1007 for communicating information.
- the bus 1007 may be composed of a single bus, or may be composed of different buses between devices.
- the behavior change promotion device 1 includes hardware such as a microprocessor, a digital signal processor (DSP), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), and an FPGA (Field Programmable Gate Array). part or all of each functional block may be implemented by the hardware.
- processor 1001 may be implemented with at least one of these hardware.
- LTE Long Term Evolution
- LTE-A Long Term Evolution-Advanced
- SUPER 3G IMT-Advanced
- 4G 5G
- FRA Full Radio Access
- W-CDMA registered trademark
- GSM registered trademark
- CDMA2000 Code Division Multiple Access 2000
- UMB Universal Mobile Broad-band
- IEEE 802.11 Wi-Fi
- IEEE 802.16 WiMAX
- IEEE 802.20 UWB (Ultra-Wide Band)
- Bluetooth® other suitable systems and/or extended next generation systems based on these.
- Input and output information may be saved in a specific location (for example, memory) or managed in a management table. Input/output information and the like can be overwritten, updated, or appended. The output information and the like may be deleted. The entered information and the like may be transmitted to another device.
- the determination may be made by a value represented by one bit (0 or 1), by a true/false value (Boolean: true or false), or by numerical comparison (for example, a predetermined value).
- notification of predetermined information is not limited to being performed explicitly, but may be performed implicitly (for example, not notifying the predetermined information). good too.
- Software whether referred to as software, firmware, middleware, microcode, hardware description language or otherwise, includes instructions, instruction sets, code, code segments, program code, programs, subprograms, and software modules. , applications, software applications, software packages, routines, subroutines, objects, executables, threads of execution, procedures, functions, and the like.
- software, instructions, etc. may be transmitted and received via a transmission medium.
- the software can be used to access websites, servers, or other When transmitted from a remote source, these wired and/or wireless technologies are included within the definition of transmission media.
- data, instructions, commands, information, signals, bits, symbols, chips, etc. may refer to voltages, currents, electromagnetic waves, magnetic fields or magnetic particles, light fields or photons, or any of these. may be represented by a combination of
- information, parameters, etc. described in this specification may be represented by absolute values, may be represented by relative values from a predetermined value, or may be represented by corresponding other information. .
- Communication terminals are defined by those skilled in the art as mobile communication terminals, subscriber stations, mobile units, subscriber units, wireless units, remote units, mobile devices, wireless devices, wireless communication devices, remote devices, mobile subscriber stations, access terminals, It may also be called a mobile terminal, wireless terminal, remote terminal, handset, user agent, mobile client, client or some other suitable term.
- any reference to the elements does not generally limit the quantity or order of those elements. These designations may be used herein as a convenient method of distinguishing between two or more elements. Thus, references to first and second elements do not imply that only two elements may be employed therein, or that the first element must precede the second element in any way.
Abstract
Description
Claims (10)
- ユーザに対して、不測の事態のリスクを低減するための情報を提示する行動変容促進装置であって、
ユーザに係る学習用ユーザ情報と、ユーザのリスクに係る情報であるリスク情報とを互いに対応付けて学習し、ユーザのリスクを推定するための第1学習モデルを構築する第1学習部と、
前記学習用ユーザ情報に含まれる情報であってリスクに影響を与えるリスク要因に係る情報と、前記リスク情報とを互いに対応付けて学習し、リスク要因を有することに伴うリスクの増加度合いであるリスク因果効果を推定するための第2学習モデルを構築する第2学習部と、
ユーザに係る推定用ユーザ情報を前記第1学習モデルに入力することにより、ユーザのリスクを推定する第1推定部と、
前記第2学習モデルを用いて前記リスク因果効果を推定する第2推定部と、
前記第1推定部によって推定されたリスク、及び、前記第2推定部によって推定された前記リスク因果効果を少なくとも含むアドバイス情報を生成するアドバイス生成部と、
前記アドバイス情報を出力する出力部と、を備える行動変容促進装置。 - 前記リスク要因に係る情報は、リスク要因である特徴量と、リスク要因又はリスクの少なくとも一方に影響を与える特徴量とを含み、
前記第2学習部は、前記リスク要因である特徴量と、前記リスク要因又はリスクの少なくとも一方に影響を与える特徴量と、前記リスク情報とを互いに対応付けて学習し、前記第2学習モデルを構築する、請求項1記載の行動変容促進装置。 - 前記リスク要因又はリスクの少なくとも一方に影響を与える特徴量は、ユーザの特徴を示す情報を含む、請求項2記載の行動変容促進装置。
- 前記第2推定部は、前記推定用ユーザ情報に含まれる情報であってリスクに影響を与えるリスク要因に係る情報に含まれる、リスク要因又はリスクの少なくとも一方に影響を与える特徴量を前記第2学習モデルに入力することにより、前記リスク因果効果を推定する、請求項1~3のいずれか一項に記載の行動変容促進装置。
- 前記第2学習モデルに基づき、前記推定用ユーザ情報に含まれる情報であってリスクに影響を与えるリスク要因に係る情報に含まれる、前記リスク要因又はリスクの少なくとも一方に影響を与える特徴量が、前記リスク因果効果に与える影響度を算出する算出部を更に備える、請求項1~4のいずれか一項に記載の行動変容促進装置。
- 前記算出部は、前記第2学習モデルに基づき算出された、前記リスク要因又はリスクの少なくとも一方に影響を与える特徴量が前記リスク因果効果に与える影響度と、ユーザにおける前記リスク要因又はリスクの少なくとも一方に影響を与える特徴量の保有の有無に係る情報とに基づき、当該ユーザが保有する特徴量が前記リスク因果効果に与える影響度を算出する、請求項5記載の行動変容促進装置。
- 前記アドバイス生成部は、前記算出部によって算出された、前記リスク因果効果に対する前記リスク要因又はリスクの少なくとも一方に影響を与える特徴量の影響度を更に含む前記アドバイス情報を生成する、請求項6記載の行動変容促進装置。
- 前記アドバイス生成部は、前記第2推定部によって推定された前記リスク因果効果に係る前記リスク要因に予め対応付けられた情報であってリスクを小さくするためにユーザに促す行動を示す情報である行動変容促進内容を更に含む前記アドバイス情報を生成する、請求項1~7のいずれか一項に記載の行動変容促進装置。
- 前記出力部は、前記アドバイス情報について、最初に前記第1推定部によって推定されたリスクを出力し、つづいて前記第2推定部によって推定された前記リスク因果効果を出力し、最後に前記行動変容促進内容を出力する、請求項8記載の行動変容促進装置。
- 前記第2学習モデルに基づき、前記推定用ユーザ情報に含まれる情報であってリスクに影響を与えるリスク要因に係る情報に含まれる、前記リスク要因又はリスクの少なくとも一方に影響を与える特徴量が、前記リスク因果効果に与える影響度を算出する算出部を更に備え、
前記出力部は、前記アドバイス情報について、前記第2推定部によって推定された前記リスク因果効果を出力した後、前記行動変容促進内容を出力する前に、前記算出部によって算出された前記リスク因果効果に対する前記リスク要因又はリスクの少なくとも一方に影響を与える特徴量の影響度を出力する、請求項9記載の行動変容促進装置。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2023512882A JP7413607B2 (ja) | 2021-04-07 | 2022-03-11 | 行動変容促進装置 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021065147 | 2021-04-07 | ||
JP2021-065147 | 2021-04-07 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022215435A1 true WO2022215435A1 (ja) | 2022-10-13 |
Family
ID=83546356
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2022/011025 WO2022215435A1 (ja) | 2021-04-07 | 2022-03-11 | 行動変容促進装置 |
Country Status (2)
Country | Link |
---|---|
JP (1) | JP7413607B2 (ja) |
WO (1) | WO2022215435A1 (ja) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2010122901A (ja) * | 2008-11-19 | 2010-06-03 | Omron Healthcare Co Ltd | 健康状態判断装置 |
JP2018124702A (ja) * | 2017-01-31 | 2018-08-09 | 株式会社教育ソフトウェア | 病因分析装置および疾病予測装置 |
US20180322955A1 (en) * | 2017-05-03 | 2018-11-08 | Koninklijke Philips N.V. | Visually indicating contributions of clinical risk factors |
JP2021039518A (ja) * | 2019-09-02 | 2021-03-11 | 株式会社Nttドコモ | 行動変容促進装置 |
-
2022
- 2022-03-11 WO PCT/JP2022/011025 patent/WO2022215435A1/ja active Application Filing
- 2022-03-11 JP JP2023512882A patent/JP7413607B2/ja active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2010122901A (ja) * | 2008-11-19 | 2010-06-03 | Omron Healthcare Co Ltd | 健康状態判断装置 |
JP2018124702A (ja) * | 2017-01-31 | 2018-08-09 | 株式会社教育ソフトウェア | 病因分析装置および疾病予測装置 |
US20180322955A1 (en) * | 2017-05-03 | 2018-11-08 | Koninklijke Philips N.V. | Visually indicating contributions of clinical risk factors |
JP2021039518A (ja) * | 2019-09-02 | 2021-03-11 | 株式会社Nttドコモ | 行動変容促進装置 |
Non-Patent Citations (1)
Title |
---|
S\"OREN R. K\"UNZEL; JASJEET S. SEKHON; PETER J. BICKEL; BIN YU: "Meta-learners for Estimating Heterogeneous Treatment Effects using Machine Learning", ARXIV.ORG, 12 June 2017 (2017-06-12), pages 1 - 43, XP081323472, DOI: 10.1073/pnas.1804597116 * |
Also Published As
Publication number | Publication date |
---|---|
JPWO2022215435A1 (ja) | 2022-10-13 |
JP7413607B2 (ja) | 2024-01-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11263550B2 (en) | Audit machine learning models against bias | |
US11790269B1 (en) | Systems and methods for generating dynamic models based on trigger events | |
US10325243B1 (en) | Systems and methods for identifying and ranking successful agents based on data analytics | |
CN109635990B (zh) | 一种训练方法、预测方法、装置、电子设备及存储介质 | |
CN104731845A (zh) | 用于背景推送通知服务的方法和系统 | |
US20150269195A1 (en) | Model updating apparatus and method | |
US20200168012A1 (en) | Personal fuel efficiency for vehicles of interest | |
CN113794683A (zh) | 一种工业物联网入侵检测方法、装置、设备及存储介质 | |
WO2021039797A1 (ja) | クリック率予測モデル構築装置 | |
JP7033515B2 (ja) | 交通状況予測装置 | |
WO2022215435A1 (ja) | 行動変容促進装置 | |
CN108628863B (zh) | 信息获取方法和装置 | |
JP6855359B2 (ja) | 渋滞予測モデル作成装置 | |
CN116862580A (zh) | 短信触达时间预测方法、装置、计算机设备及存储介质 | |
CN110704614A (zh) | 对应用中的用户群类型进行预测的信息处理方法及装置 | |
JP2019220100A (ja) | 推定装置 | |
CN113794682B (zh) | 一种工业物联网入侵检测智能体训练方法、装置及设备 | |
CN113783841B (zh) | 工业物联网入侵检测网络架构构建方法、装置、设备及存储介质 | |
WO2023013475A1 (ja) | 行動変容促進装置 | |
JPWO2019159585A1 (ja) | 学習システム、推定システム及び学習済モデル | |
CN113807436A (zh) | 用户挖掘方法、装置、计算机设备及可读存储介质 | |
JP6835680B2 (ja) | 情報処理装置および信用度算出方法 | |
WO2023067895A1 (ja) | 情報処理装置 | |
CN111127040A (zh) | 数据处理的方法、智能客服系统、电子设备和介质 | |
CN111178687B (zh) | 金融风险分类方法、装置及电子设备 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22784420 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2023512882 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 18553587 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 22784420 Country of ref document: EP Kind code of ref document: A1 |