Information processing device, information processing method, and computer program product

An information processing device comprising

• one or more hardware processors configured to:
• cause a display device to display a time-series change in time-series target sensor data obtained by observing an observation target before a base time;
• and
• cause the display device to display a probability distribution of a maximum value or a minimum value of the target sensor data in an estimation period after the base time.

An information processing method implemented by a computer, the method being for estimating a probability distribution of a maximum value or a minimum value of time-series target sensor data in a designated estimation period after a base time, the time-series target data being obtained by observing an observation target, the method comprising:

• calculating a reference feature amount representing a feature of one or more pieces of related sensor data at a reference time before the base time, the reference feature amount being calculated based on:
• one or more pieces of time-series related sensor data obtained by observing one or more related targets observed together with the observation target before the reference time;
• and a preset feature amount calculation model;
• estimating a reference distribution parameter being a distribution parameter of a probability density function representing the probability distribution of the maximum value or the minimum value of the target sensor data in a reference period in which a time relation with respect to the reference time is the same as a time relation of the estimation period with respect to the base time, the reference distribution parameter being estimated based on the reference feature amount at the reference time and a preset parameter estimation model;
• updating at least one of the feature amount calculation model and the parameter estimation model based on:
• an actual measurement value of the maximum value or the minimum value of the target sensor data in the reference period corresponding to the reference time;
• and the probability distribution of the maximum value or the minimum value of the target sensor data in the reference period, the probability distribution being generated for the reference time based on the reference distribution parameter;
• calculating an estimation feature amount representing the feature of the one or more pieces of related sensor data at the base time, the estimation feature amount being calculated based on:
• the one or more pieces of time-series related sensor data before the base time;
• and the updated feature amount calculation model;
• and
• estimating an estimation distribution parameter being the distribution parameter of the probability density function representing the probability distribution of the maximum value or the minimum value of the target sensor data in the estimation period, the estimation distribution parameter being estimated based on the estimation feature amount at the base time and the updated parameter estimation model.

A computer program product comprising a non-transitory computer-readable recording medium on which an executable program for a computer is recorded, the program instructing the computer to estimate a probability distribution of a maximum value or a minimum value of time-series target sensor data in a designated estimation period after a base time, the time-series target data being obtained by observing an observation target, the program further instructing the computer to:

• calculate a reference feature amount representing a feature of one or more pieces of related sensor data at a reference time before the base time, the reference feature amount being calculated based on:
• one or more pieces of time-series related sensor data obtained by observing one or more related targets observed together with the observation target before the reference time;
• and a preset feature amount calculation model;
• estimate a reference distribution parameter being a distribution parameter of a probability density function representing the probability distribution of the maximum value or the minimum value of the target sensor data in a reference period in which a time relation with respect to the reference time is the same as a time relation of the estimation period with respect to the base time, the reference distribution parameter being estimated based on the reference feature amount at the reference time and a preset parameter estimation model;
• update at least one of the feature amount calculation model and the parameter estimation model based on:
• an actual measurement value of the maximum value or the minimum value of the target sensor data in the reference period corresponding to the reference time;
• and the probability distribution of the maximum value or the minimum value of the target sensor data in the reference period, the probability distribution being generated for the reference time based on the reference distribution parameter;
• calculate an estimation feature amount representing the feature of the one or more pieces of related sensor data at the base time, the estimation feature amount being calculated based on:
• the one or more pieces of time-series related sensor data before the base time;
• and the updated feature amount calculation model;
• and
• estimate an estimation distribution parameter being the distribution parameter of the probability density function representing the probability distribution of the maximum value or the minimum value of the target sensor data in the estimation period, the estimation distribution parameter being estimated based on the estimation feature amount at the base time and the updated parameter estimation model.