Abstract: A method includes: partitioning learning data containing objective variables and explanatory variables into a plurality of subsets of data; executing regularization processing on first data in each of the partitioned subsets, and extracting a first element equal to zero; extracting, as a candidate, each model where an error ratio between first multiple regression and second multiple regression is equal to or more than a predetermined value, the first multiple regression being a result of multiple regression on second data which is test data in each of the partitioned subsets and is for use to calculate the error ratio of the learning data, the second multiple regression being a result of multiple regression on third data obtained by excluding the first element from the second data; and outputting a model where zero is substituted for an element that takes zero a predetermined or larger number of times in the candidate.

Abstract: A method includes: partitioning learning data containing objective variables and explanatory variables into a plurality of subsets of data; executing regularization processing on first data in each of the partitioned subsets, and extracting a first element equal to zero; extracting, as a candidate, each model where an error ratio between first multiple regression and second multiple regression is equal to or more than a predetermined value, the first multiple regression being a result of multiple regression on second data which is test data in each of the partitioned subsets and is for use to calculate the error ratio of the learning data, the second multiple regression being a result of multiple regression on third data obtained by excluding the first element from the second data; and outputting a model where zero is substituted for an element that takes zero a predetermined or larger number of times in the candidate.

Abstract: A search device, for each of a plurality of inquiries including a plurality of options, based on a log of answers of a respondent in the past, specifies a matching option corresponding to a trend of answers of the respondent from the plurality of options to be represented to the respondent. The search device, when an answer to the plurality of options is included in the matching option, determines that it is a preference-reached mode that is a mode where the estimating a preference of the user converges. The search device transmits, to the respondent, an inquiry including options, contents of the options included in the inquiry being changed depend on whether it is determined to be a preference-reached mode in the determining.

Abstract: A method includes: partitioning learning data containing objective variables and explanatory variables into a plurality of subsets of data; executing regularization processing on first data in each of the partitioned subsets, and extracting a first element equal to zero; extracting, as a candidate, each model where an error ratio between first multiple regression and second multiple regression is equal to or more than a predetermined value, the first multiple regression being a result of multiple regression on second data which is test data in each of the partitioned subsets and is for use to calculate the error ratio of the learning data, the second multiple regression being a result of multiple regression on third data obtained by excluding the first element from the second data; and outputting a model where zero is substituted for an element that takes zero a predetermined or larger number of times in the candidate.

Abstract: A search device generates a preference trend based on content of a plurality of answers from a respondent for a plurality of inquiries each including a plurality of options. The search device, based on a log of content of representation items that are represented at the plurality of inquiries to the respondent, determines whether information that changes the representation items is to be represented. The search device generates an input candidate that changes the representation items according to a determining result based on the preference trend and displays the input candidate together with the representation items at the plurality of inquiries.

Abstract: A search device, for each of a plurality of inquiries including a plurality of options, based on a log of answers of a respondent in the past, specifies a matching option corresponding to a trend of answers of the respondent from the plurality of options to be represented to the respondent. The search device, when an answer to the plurality of options is included in the matching option, determines that it is a preference-reached mode that is a mode where the estimating a preference of the user converges. The search device transmits, to the respondent, an inquiry including options, contents of the options included in the inquiry being changed depend on whether it is determined to be a preference-reached mode in the determining.

Abstract: An information processing unit and information processing method capable of performing a ground-state search of an Ising model having coefficients of arbitrary values regardless of restrictions on hardware or software is suggested. When a ground state of an original problem, which is an Ising model, or an approximate solution of that ground state is calculated as a solution of the original problem, one or more sub-problems which are Ising models are generated from the original problem and the information processing unit searches the ground state of each generated sub-problem and generates a solution of the original problem based on a solution of each sub-problem obtained from the search; and when types of coefficient values of an Ising model whose ground state can be searched are limited, sub-problems which are Ising models composed of the types of limited values of coefficients are generated.

Abstract: A search apparatus specifies a first respondent group based on a respondent attribute of a respondent. The search apparatus searches for a respondent preference of the respondent utilizing a plurality of inquiries for the respondent, the plurality of inquiries being made with a first preference of the first respondent group as an initial value and corresponding to a first segment that the first respondent group belongs to. The search apparatus, when a specific condition is not satisfied after search from the initial value with a specific number of inquiries, generates or selects a second respondents group based on the respondent attribute, answers to the plurality of inquiries, and on preferences of a second respondent group, the second respondent group satisfying the specific condition. The search apparatus searches for the respondent preference with a preference of the second respondent group being an initial value.

Abstract: An information processing apparatus executes 1 learning on each of a plurality of neural networks with regard to target data by duration of at least 1 epoch, and information processing apparatus executes a plurality times of loops of a specific algorithm, each of the plurality of times of loops changes a number of units of each of the plurality of neural networks. The information processing apparatus sets a plurality of learning durations for each of the plurality of neural networks based on a respective accuracy variance and a respective actual performance.

Abstract: An information processing unit and information processing method capable of performing a ground-state search of an Ising model having coefficients of arbitrary values regardless of restrictions on hardware or software is suggested. When a ground state of an original problem, which is an Ising model, or an approximate solution of that ground state is calculated as a solution of the original problem, one or more sub-problems which are Ising models are generated from the original problem and the information processing unit searches the ground state of each generated sub-problem and generates a solution of the original problem based on a solution of each sub-problem obtained from the search; and when types of coefficient values of an Ising model whose ground state can be searched are limited, sub-problems which are Ising models composed of the types of limited values of coefficients are generated.