Relationship analysis device, relationship analysis method, and recording medium for analyzing relationship between a plurality of types of data using kernel mean learning
A relationship analysis device includes a parameter sample data calculation unit that calculates sample data for parameters for a simulator that receives inputs of data of a first type and outputs data of a second type, calculating sample data; a second type sample data acquisition unit that inputs, to the simulator, observation data and sample data, and obtains sample data of the second type; and a parameter value determination unit that calculates a weight for sample data based on the difference between observation data of the second type and the sample data of the second type, and based on the relationship between a first distribution that the observation data of the first type followed and a second distribution being a distribution of the data of the first type, and calculates a value for the parameters using the calculated weight.
1. A relationship analysis device for training a machine learning model, the relationship analysis device comprising:
at least one memory configured to store instructions; and
at least one processor configured to execute the instructions to:
calculate a plurality of pieces of sample data for parameters for a simulator, based on a temporarily set distribution for the parameters, the simulator including a non-differentiable function, receiving inputs of data of a first type and outputting data of a second type;
input, to the simulator, observation data of the first type and each of the plurality of pieces of sample data for the parameters, and obtain a sample data of the second type for each of the plurality of pieces of sample data for the parameters;
calculate a weight for each of the plurality of pieces of sample data for the parameters based on a difference between observation data of the second type and the sample data of the second type, and based on the relationship between a first distribution that the observation data of the first type followed and a second distribution being a distribution of the data of the first type and indicating a region for which a relationship is to be found, wherein the region indicated by the second distribution is used to improve accuracy in response to a covariate shift in the data of the first type;
calculate a value for the parameters using the calculated weight;
calculate a kernel mean in which the degree of agreement of each element of the data of the first type with the second distribution is reflected in a posterior distribution of the parameters under the observation data of the first type and the calculated sample data of the second type;
recursively calculate sample data of the parameters based on the kernel mean using kernel herding method; and
present the observation data of the first type, the observation data of the second type, and the calculated sample data on a display to a user.
2. The relationship analysis device according to claim 1 , wherein the at least one processor is configured to execute the instructions to:
calculate a kernel expression of the predictive distribution of the parameters using sample data of the parameters based on the kernel mean; and
calculate sample data according to the predictive distribution of the data of the second type by using the kernel expression of the predictive distribution of the parameters.
3. A relationship analysis method for training a machine learning model, the relationship analysis method comprising:
calculating a plurality of pieces of sample data for parameters for a simulator that includes a non-differentiable function, receives inputs of data of a first type and outputs data of a second type, calculating the of pieces of sample data based on a temporarily set distribution for the parameters;
inputting, to the simulator, observation data of the first type and a plurality of pieces of sample data for the parameters, and calculating sample data of the second type for each of the plurality of pieces of sample data for the parameters;
calculating a weight for each of the plurality of pieces of sample data for the parameters based on a difference between observation data of the second type and the sample data of the second type, and based on the relationship between a first distribution that the observation data of the first type followed and a second distribution being a distribution of the data of the first type and indicating a region for which a relationship is to be found, wherein the region indicated by the second distribution is used to improve accuracy in response to a covariate shift in the data of the first type;
calculating a value for the parameters using the calculated weight;
calculating a kernel mean in which the degree of agreement of each element of the data of the first type with the second distribution is reflected in a posterior distribution of the parameters under the observation data of the first type and the calculated sample data of the second type;
recursively calculating sample data of the parameters based on the kernel mean using kernel herding method; and
presenting the observation data of the first type, the observation data of the second type, and the calculated sample data on a display to a user.
4. A non-transitory recording medium that records, for training a machine learning model, a program for causing a computer to execute:
calculating a plurality of pieces of sample data for parameters for a simulator that includes non-differentiable function, receives inputs of data of a first type and outputs data of a second type, calculating the plurality of pieces of sample data based on a temporarily set distribution for the parameters;
inputting, to the simulator, observation data of the first type and a plurality of pieces of sample data for the parameters, and calculating sample data of the second type for each of the plurality of pieces of sample data for the parameters;
calculating a weight for each of the plurality of pieces of sample data for the parameters based on a difference between observation data of the second type and the sample data of the second type, and based on the relationship between a first distribution that the observation data of the first type followed and a second distribution being a distribution of the data of the first type and indicating a region for which a relationship is to be found, wherein the region indicated by the second distribution is used to improve accuracy in response to a covariate shift in the data of the first type;
calculating a value for the parameters using the calculated weight;
calculating a kernel mean in which the degree of agreement of each element of the data of the first type with the second distribution is reflected in a posterior distribution of the parameters under the observation data of the first type and the calculated sample data of the second type;
recursively calculating sample data of the parameters based on the kernel mean using kernel herding method; and
presenting the observation data of the first type, the observation data of the second type, and the calculated sample data on a display to a user.
5. The relationship analysis device according to claim 1 , wherein the at least one processor is configured to execute the instructions to estimate distribution of the parameters based on the recursively calculated sample data.
6. The relationship analysis device for training the machine learning model according to claim 1 , wherein the calculating the weight includes calculating the parameters in the machine learning model.