Information processing device, information processing method, and program

Abstract

If there are missing parts in the observation data acquired from the sensor, these missing parts are filled in. [Solution] The information processing device 10 includes a pre-setting unit 11 that sets initial values ​​for missing portions in observation data output from a sensor, a simulation data evaluation unit 12 that acquires the results of multiple simulations performed assuming an observation environment and evaluates the similarity between the observation data including missing portions and the data obtained in each simulation, and a missing value estimation unit 13 that estimates the value of the missing portion using the data obtained in the simulation with the highest similarity.

Claims

1. A pre-setting unit sets the initial values ​​for missing portions in the observation data output from the sensor, A simulation data evaluation unit acquires the results of multiple simulations performed under assumed observation environments, and evaluates the similarity between the observation data, including the missing portions, and the data obtained from each simulation. A missing value estimation unit estimates the value of the missing portion using the data obtained from the simulation that has the highest similarity, An information processing device characterized by having the following features.

2. The pre-setting unit sets the initial value by setting a prior distribution whose probability changes according to the value, The simulation data evaluation unit performs the evaluation by calculating a similarity score that indicates the similarity between the observed data values, including the values ​​obtained from the prior distribution of the missing portion, and the data obtained from the simulation. The information processing apparatus according to claim 1.

3. The pre-setting unit sets the prior distribution such that the probability of the observed data being between a lower limit and a first value is constant, and the probability of it being between the first value and an upper limit is decreased. The information processing apparatus according to claim 2.

4. The pre-setting unit sets the prior distribution such that the probability is maximized when the observed data is the same value as a measured value obtained by a sensor other than the sensor, the probability is increased from the lower limit to the same value, and the probability is decreased from the same value to the upper limit. The information processing apparatus according to claim 2.

5. The missing value estimation unit replaces the missing portion of the observed data with the estimated value. The information processing apparatus according to claim 1.

6. The sensor is a sensing device that uses optical fiber cables laid in the road. The aforementioned observation data is average speed data, which represents the average speed of vehicles traveling on the road, calculated using data from the optical fiber cable. The pre-setting unit sets a prior distribution whose probability changes according to the average speed of the vehicle. The information processing apparatus according to claim 2.

7. A pre-configuration step to set initial values ​​for missing portions in the observation data output from the sensor, A simulation data evaluation step involves obtaining the results of multiple simulations performed under assumed observation environments, and for each simulation, evaluating the similarity between the observation data, including the missing portions, and the data obtained from that simulation. A missing value estimation step, in which the value of the missing portion is estimated using the data obtained from the simulation that has the highest similarity, An information processing method characterized by having the following:

8. In the aforementioned pre-setting step, the initial value is set by setting a prior distribution whose probability changes depending on the value, The simulation data evaluation step performs the evaluation by calculating a similarity score that indicates the similarity between the observed data values, including the values ​​obtained from the prior distribution of the missing portion, and the data obtained from the simulation. The information processing method according to claim 7.

9. In the aforementioned pre-setting step, the prior distribution is set such that the probability of the observed data falling from the lower limit to a first value is constant, and the probability of falling from the first value to the upper limit is decreased. The information processing method according to claim 8.

10. In the aforementioned pre-setting step, the prior distribution is set such that the probability is maximized when the observed data is the same value as a measured value obtained from a sensor other than the aforementioned sensor, the probability is increased from the lower limit to the same value, and the probability is decreased from the same value to the upper limit. The information processing method according to claim 8.

11. In the missing value estimation step, the missing portion of the observed data is replaced with the estimated value. The information processing method according to claim 7.

12. The sensor is a sensing device that uses optical fiber cables laid in the road. The aforementioned observation data is average speed data, which represents the average speed of vehicles traveling on the road, calculated using data from the optical fiber cable. In the aforementioned pre-setting step, a prior distribution is set whose probability changes according to the average speed of the vehicle. The information processing method according to claim 8.

13. On the computer, A pre-configuration step to set initial values ​​for missing portions in the observation data output from the sensor, A simulation data evaluation step involves obtaining the results of multiple simulations performed under assumed observation environments, and for each simulation, evaluating the similarity between the observation data, including the missing portions, and the data obtained from that simulation. A missing value estimation step, in which the value of the missing portion is estimated using the data obtained from the simulation that has the highest similarity, A program that executes something.

14. In the aforementioned pre-setting step, the initial value is set by setting a prior distribution whose probability changes depending on the value, The simulation data evaluation step performs the evaluation by calculating a similarity score that indicates the similarity between the observed data values, including the values ​​obtained from the prior distribution of the missing portion, and the data obtained from the simulation. The program according to claim 13.

15. In the aforementioned pre-setting step, the prior distribution is set such that the probability of the observed data falling from the lower limit to a first value is constant, and the probability of falling from the first value to the upper limit is decreased. The program according to claim 14.

16. In the aforementioned pre-setting step, the prior distribution is set such that the probability is maximized when the observed data is the same value as a measured value obtained from a sensor other than the aforementioned sensor, the probability is increased from the lower limit to the same value, and the probability is decreased from the same value to the upper limit. The program according to claim 14.

17. In the missing value estimation step, the missing portion of the observed data is replaced with the estimated value. The program according to claim 13.

18. The sensor is a sensing device that uses optical fiber cables laid in the road. The aforementioned observation data is average speed data, which represents the average speed of vehicles traveling on the road, calculated using data from the optical fiber cable. In the aforementioned pre-setting step, a prior distribution is set whose probability changes according to the average speed of the vehicle. The program according to claim 14.

Images