Patents by Inventor Yasue Makino
Yasue Makino has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Publication number: 20220374748Abstract: A determination is made of an explanatory variable with respect to an objective variable. A subset of data from data to be analyzed is created, in response to setting the objective variable to be analyzed to perform analysis. Association analysis is applied to analysis results, in response to a number of analysis runs exceeding a predetermined number. An association rule is derived for the explanatory variable from a result of the association analysis. An explanatory variable having a relevance value greater than a threshold value with the objective variable in the data to be analyzed is selected. The selected explanatory variable is scored as an input using the association rule to determine whether the explanatory variable is to be added or removed.Type: ApplicationFiled: August 2, 2022Publication date: November 24, 2022Inventors: Hiromi KOBAYASHI, Masaharu SAKAMOTO, Yasue MAKINO, Hirokazu KOBAYASHI
-
Patent number: 11410064Abstract: A determination is made of an explanatory variable with respect to an objective variable. A subset of data from data to be analyzed is created, in response to setting the objective variable to be analyzed to perform analysis. Association analysis is applied to analysis results, in response to a number of analysis runs exceeding a predetermined number. An association rule is derived for the explanatory variable from a result of the association analysis. An explanatory variable having a relevance value greater than a threshold value with the objective variable in the data to be analyzed is selected. The selected explanatory variable is scored as an input using the association rule to determine whether the explanatory variable is to be added or removed.Type: GrantFiled: January 14, 2020Date of Patent: August 9, 2022Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Hiromi Kobayashi, Masaharu Sakamoto, Yasue Makino, Hirokazu Kobayashi
-
Patent number: 11295177Abstract: In an approach to improving accuracy through weak model aggregation, one or more computer processors generating a plurality of hyperparameter sets, wherein each hyperparameter set in the plurality of hyperparameter sets contains one or more hyperparameters varied to increase over-training in one or more models, wherein over-training includes overfitting or underfitting. The one or more computer processors create a plurality of weak models utilizing a created bootstrap dataset in a plurality of created bootstrap datasets, a corresponding extracted explanatory variable set, and a corresponding hyperparameter set in the generated plurality of hyperparameter sets, wherein each weak model in a created plurality of weak models shares at least the created bootstrap dataset, the extracted explanatory variable set, the generated hyperparameter set, a machine learning technique, or a model architecture.Type: GrantFiled: March 27, 2020Date of Patent: April 5, 2022Assignee: International Business Machines CorporationInventors: Masaharu Sakamoto, Yasue Makino, Hiromi Kobayashi, Hirokazu Kobayashi
-
Patent number: 11270321Abstract: Reducing noise during association analysis is provided. An association model is generated having a support value with respect to original data greater than a predefined minimum support value threshold level. A number of association rules corresponding to the association model are identified. It is determined whether the number of association rules corresponding to the association model is less than a predefined maximum number of association rules. In response to determining that the number of association rules corresponding to the association model is greater than the predefined maximum number of association rules, inverted data of the original data is generated. Another association model is generated having the support value with respect to the original data and the inverted data greater than the predefined minimum support value threshold level.Type: GrantFiled: August 27, 2019Date of Patent: March 8, 2022Assignee: International Business Machines CorporationInventors: Yusuke Matsumoto, Yasue Makino, Hirokazu Kobayashi, Hiromi Kobayashi
-
Publication number: 20210303937Abstract: In an approach to improving accuracy through weak model aggregation, one or more computer processors generating a plurality of hyperparameter sets, wherein each hyperparameter set in the plurality of hyperparameter sets contains one or more hyperparameters varied to increase over-training in one or more models, wherein over-training includes overfitting or underfitting. The one or more computer processors create a plurality of weak models utilizing a created bootstrap dataset in a plurality of created bootstrap datasets, a corresponding extracted explanatory variable set, and a corresponding hyperparameter set in the generated plurality of hyperparameter sets, wherein each weak model in a created plurality of weak models shares at least the created bootstrap dataset, the extracted explanatory variable set, the generated hyperparameter set, a machine learning technique, or a model architecture.Type: ApplicationFiled: March 27, 2020Publication date: September 30, 2021Inventors: Masaharu Sakamoto, YASUE MAKINO, HIROMI KOBAYASHI, HIROKAZU KOBAYASHI
-
Publication number: 20210216895Abstract: A determination is made of an explanatory variable with respect to an objective variable. A subset of data from data to be analyzed is created, in response to setting the objective variable to be analyzed to perform analysis. Association analysis is applied to analysis results, in response to a number of analysis runs exceeding a predetermined number. An association rule is derived for the explanatory variable from a result of the association analysis. An explanatory variable having a relevance value greater than a threshold value with the objective variable in the data to be analyzed is selected. The selected explanatory variable is scored as an input using the association rule to determine whether the explanatory variable is to be added or removed.Type: ApplicationFiled: January 14, 2020Publication date: July 15, 2021Inventors: Hiromi KOBAYASHI, Masaharu SAKAMOTO, Yasue MAKINO, Hirokazu KOBAYASHI
-
Publication number: 20210065214Abstract: Reducing noise during association analysis is provided. An association model is generated having a support value with respect to original data greater than a predefined minimum support value threshold level. A number of association rules corresponding to the association model are identified. It is determined whether the number of association rules corresponding to the association model is less than a predefined maximum number of association rules. In response to determining that the number of association rules corresponding to the association model is greater than the predefined maximum number of association rules, inverted data of the original data is generated. Another association model is generated having the support value with respect to the original data and the inverted data greater than the predefined minimum support value threshold level.Type: ApplicationFiled: August 27, 2019Publication date: March 4, 2021Inventors: Yusuke Matsumoto, Yasue Makino, Hirokazu Kobayashi, Hiromi Kobayashi
-
Patent number: 10891354Abstract: A method for performing responses based on time-series data includes receiving time-series data for each variable of a plurality of explanatory variables, determining a least common period for the plurality of explanatory variables, forming a plurality of explanatory variable groups using lags less than the least common period, determining one or more selected explanatory variable groups of the plurality of explanatory variable groups based on at least one information metric, receiving categorical data for a response variable corresponding to the plurality of explanatory variables, generating a response variable model using the selected explanatory variable groups, receiving additional time-series data corresponding to the selected explanatory variable groups, generating a categorization sequence for the additional time-series data using the response variable model, and performing one or more responses based on the categorization sequence.Type: GrantFiled: May 23, 2018Date of Patent: January 12, 2021Assignee: International Business Machines CorporationInventors: Yasue Makino, Hiromi Kobayashi, Yusuke Matsumoto, Hirokazu Kobayashi
-
Publication number: 20200394527Abstract: A prediction model can be created by reading an objective variable and a plurality of explanatory variables into a memory, and subsequently calculating a degree of influence of each of the plurality of explanatory variables on the objective variable. The determination of whether the two highest degrees of influence are approximate with each other or not can be performed. In the case that the two highest degrees of influence are approximate with each other, then a stepwise method can be carried out in order to select one explanatory variable among the plurality of explanatory variables. A prediction model can be subsequently created using the selected explanatory variable.Type: ApplicationFiled: June 12, 2019Publication date: December 17, 2020Inventors: Yasue Makino, Hiromi Kobayashi, Yusuke Matsumoto, Hirokazu Kobayashi
-
Publication number: 20200394528Abstract: A prediction model can be created by reading an objective variable and a plurality of explanatory variables into a memory, and subsequently calculating a degree of influence of each of the plurality of explanatory variables on the objective variable. The determination of whether the two highest degrees of influence are approximate with each other or not can be performed. In the case that the two highest degrees of influence are approximate with each other, then a stepwise method can be carried out in order to select one explanatory variable among the plurality of explanatory variables. A prediction model can be subsequently created using the selected explanatory variable.Type: ApplicationFiled: July 11, 2019Publication date: December 17, 2020Inventors: Yasue Makino, Hiromi Kobayashi, Yusuke Matsumoto, Hirokazu Kobayashi
-
Publication number: 20200218994Abstract: Methods, computer program products, and systems are presented.Type: ApplicationFiled: January 8, 2019Publication date: July 9, 2020Inventors: Hiromi KOBAYASHI, Yasue MAKINO, Yusuke MATSUMOTO, Hirokazu KOBAYASHI
-
Publication number: 20190361956Abstract: A method for performing responses based on time-series data includes receiving time-series data for each variable of a plurality of explanatory variables, determining a least common period for the plurality of explanatory variables, forming a plurality of explanatory variable groups using lags less than the least common period, determining one or more selected explanatory variable groups of the plurality of explanatory variable groups based on at least one information metric, receiving categorical data for a response variable corresponding to the plurality of explanatory variables, generating a response variable model using the selected explanatory variable groups, receiving additional time-series data corresponding to the selected explanatory variable groups, generating a categorization sequence for the additional time-series data using the response variable model, and performing one or more responses based on the categorization sequence.Type: ApplicationFiled: May 23, 2018Publication date: November 28, 2019Inventors: Yasue Makino, Hiromi Kobayashi, Yusuke Matsumoto, Hirokazu Kobayashi