US 12,136,023 B2
Information processing apparatus, information processing method, and computer program
Arika Fukushima, Ota Tokyo (JP); and Myungsook Ko, Ota Tokyo (JP)
Assigned to Kabushiki Kaisha Toshiba, Tokyo (JP)
Filed by KABUSHIKI KAISHA TOSHIBA, Tokyo (JP)
Filed on Sep. 9, 2020, as Appl. No. 17/015,824.
Claims priority of application No. 2020-044457 (JP), filed on Mar. 13, 2020.
Prior Publication US 2021/0287134 A1, Sep. 16, 2021
Int. Cl. G06N 20/00 (2019.01); G06F 17/18 (2006.01)
CPC G06N 20/00 (2019.01) [G06F 17/18 (2013.01)] 16 Claims
OG exemplary drawing
 
1. An information processing apparatus, comprising:
a difference calculator implemented by processing circuitry, configured to calculate, based on a plurality of data items each including an objective variable and explanatory variables, a plurality of differences between the objective variable and prediction values of the objective variable predicted from the explanatory-variables;
a candidate generator implemented by processing circuitry, configured to generate at least one first division condition candidate which divides each of the explanatory variables into a plurality of sections;
a division condition selector implemented by processing circuitry, configured to select a first division condition from the at least one first division condition candidate, based on the differences calculated by the difference calculator, the explanatory variable divided by the first division condition corresponding to a first explanatory variable of the explanatory variables;
a selection condition generator implemented by processing circuitry, configured to determine whether each of the sections obtained by dividing the first explanatory variable with the first division condition satisfies an anomaly condition based on a number of data items belonging to each of the sections to select a first section which is the section not satisfying the anomaly condition, and generate a first selection condition which selects data items belonging to the first section based on a range of values of the first explanatory variable specified by the first section; and
a data selector implemented by processing circuitry, configured to select first data items from among the plurality of data items, based on the first selection condition,
wherein
the candidate generator is configured to generate at least one second division condition candidate which divides each of the explanatory variables into a plurality of sections,
the division condition selector is configured to select a second division condition from the at least one second division condition candidate based on differences calculated by the difference calculator from the first data items belonging to each of the sections into which each explanatory variable is divided, the explanatory variable divided by the second division condition corresponding to a second explanatory variable different from the first explanatory variable among the explanatory variables, and
the selection condition generator is configured to determine whether each of the sections obtained by dividing the second explanatory variable with the second division condition satisfies an anomaly condition based on a number of first data items belonging to each of the sections to select a second section which is the section not satisfying the anomaly condition, and generate a second selection condition which selects data items belonging to the second section based on a range of values of the second explanatory variable specified by the second section.