Prediction device, prediction method, and non-transitory computer readable medium

Patent metadata

Field	Value
Publication number	US-2020293967-A1
Application number	US-201916567604-A
Country	US
Kind code	A1
Filing date	Sep 11, 2019
Priority date	Mar 12, 2019
Publication date	Sep 17, 2020
Grant date	—

Abstract

According to one embodiment, a prediction device includes: first processing circuitry and second processing circuitry. The first processing circuitry corrects a demand result value on a particular date on a basis of an adjustment coefficient depending on the particular date; and generates a prediction model of a demand on a basis of the corrected demand result value on the particular date and a demand result value on a non-particular date. The second processing circuitry calculates a demand predicted value on a date of a prediction target on a basis of demand result values on one or more dates, an adjustment coefficient depending on a particular date among the one or more dates, and the prediction model. The second processing circuitry inversely corrects the demand predicted value using an adjustment coefficient depending on the date of the prediction target if the date of the prediction target is a particular date.

First claim

1 . A prediction device comprising:
first processing circuitry configured to:
correct a demand result value on a particular date on a basis of an adjustment coefficient depending on the particular date; and
generate a prediction model of a demand on a basis of the corrected demand result value on the particular date and a demand result value on a non-particular date; and
second processing circuitry configured to:
calculate a demand predicted value on a date of a prediction target on a basis of demand result values on one or more dates, an adjustment coefficient depending on a particular date among the one or more dates, and the prediction model; and
inversely correct the demand predicted value using an adjustment coefficient depending on the date of the prediction target if the date of the prediction target is a particular date.

2 . The prediction device according to claim 1 , wherein,
the first circuitry is configured to:
calculate an adjustment coefficient for a particular date based on a difference between a demand result value on the particular date and a demand result value on a non-particular date;
generate adjustment coefficient data in which a plurality of the adjustment coefficients are associated with a plurality of the particular dates,
the first circuitry is configured to specify the adjustment coefficient depending on the particular date on a basis of the adjustment coefficient data, and the second circuitry is configured to specify the adjustment coefficient depending on the particular date on a basis of the adjustment coefficient data.

3 . The prediction device according to claim 2 , wherein
the first processing circuitry is configured:
calculate a number of elapsed days in a year for a plurality of the particular dates, and
generate the adjustment coefficient data in which a plurality of pieces of the number of elapsed days are associate with a plurality of the adjustment coefficients.

4 . The prediction device according to claim 3 , wherein
the first processing circuitry is configured to:
specify an adjustment coefficient corresponding to a number of elapsed days in a year closest to the particular date in the adjustment coefficient data; and
set the specified adjustment coefficient as the adjustment coefficient depending on the particular date, and
the second processing circuitry is configured to:
specify an adjustment coefficient corresponding to a number of elapsed days in a year closest to the date of the prediction target in the adjustment coefficient data; and
set the specified adjustment coefficient as the adjustment coefficient depending on the particular date.

5 . The prediction device according to claim 2 , wherein
the first processing circuitry is configured to:
calculate a mean value of a first demand result value and a second demand result value,
the first demand result value being a demand result value on a non-particular date among a plurality of days before the particular date and
the second demand result value being a demand result value on a non-particular date among a plurality of days after the particular date; and

calculate the adjustment coefficient for the particular date on a basis of a ratio between the demand result value on the particular date and the mean value.

6 . The prediction device according to claim 1 , wherein
the second circuitry is configured to:
specify a particular date among the one or more dates;
correct a demand result value on the specified particular date with an adjustment coefficient depending on the specified particular date; and
calculate the demand predicted value on the date of the prediction target on a basis of the corrected demand result value on the particular date, a demand result value on a non-particular date among the one or more dates, and the prediction model.

7 . The prediction device according to claim 1 , wherein
the first processing circuitry is configured to generate the prediction model using a weather predicted value on the particular date and a weather predicted value on the non-particular date and the first processing circuitry is configured to calculate the demand predicted value using a weather predicted value on the one or more dates.

8 . The prediction device according to claim 1 , wherein
the first processing circuitry is configured to correct a demand result value at first time on the particular date on a basis of an adjustment coefficient depending on the first time on the particular date, the first processing circuitry is configured to generate a prediction model of a demand at the first time on the date of the prediction target on a basis of the corrected demand result value at the first time on the particular date and the demand result value at the first time on the non-particular date, the second processing circuitry calculates a demand predicted value at the first time on the date of the prediction target on a basis of: demand result values at the first time on one or more dates before the first time on the date of the prediction target; an adjustment coefficient depending on the first time on the particular date among the one or more dates; and the prediction model, and the second processing circuitry is configured to inversely correct the demand predicted value using an adjustment coefficient depending on the first time on the date of the prediction target if the date of the prediction target is a particular date.

9 . The prediction device according to claim 1 , wherein
the first processing circuitry is configured to correct the demand result value by dividing the demand result value by the adjustment coefficient, and the second processing circuitry is configured to inversely correct the demand predicted value by multiplying the demand predicted value by the adjustment coefficient.

10 . The prediction device according to claim 1 , wherein
the demand result value is a result value of an electric power supply amount, and the demand predicted value is a predicted value of the electric power supply amount.

11 . The prediction device according to claim 1 , wherein
the prediction model is a regression model.

12 . The prediction device according to claim 1 , further comprising:
an output device configured to display the demand predicted value which is inversely corrected by the post-processor.

13 . A prediction method comprising:
correcting a demand result value on a particular date on a basis of an adjustment coefficient depending on the particular date; generating a prediction model of a demand on a basis of the corrected demand result value on the particular date and a demand result value on a non-particular date; calculating a demand predicted value on a date of a prediction target on a basis of demand result values on one or more dates, an adjustment coefficient depending on a particular date among the one or more dates, and the prediction model; and inversely correcting the demand predicted value using an adjustment coefficient depending on the date of the prediction target if the date of the prediction target is a particular date.

14 . A non-transitory computer readable medium having a computer program stored therein which causes a computer to perform processes comprising:
correcting a demand result value on a particular date on a basis of an adjustment coefficient depending on the particular date; generating a prediction model of a demand on a basis of the corrected demand result value on the particular date and a demand result value on a non-particular date; calculating a demand predicted value on a date of a prediction target on a basis of

Assignees

• Toshiba Kk
• Toshiba Energy Systems & Solutions Corp

Inventors

• Shin Hiromasa
• Shiga Yoshiaki
• Ichikawa Ryoichi
• Toyoshima Lchiro

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N20/00

  Machine learning · CPC title

• G06N3/08

  Learning methods · CPC title

• G06Q10/06314 Primary

  Calendaring for a resource · CPC title

• G06Q10/109

  Time management, e.g. calendars, reminders, meetings or time accounting · CPC title