Data analysis apparatus and method
AssigneeKabushiki Kaisha Toshiba
InventorsTakayuki ITOH, Wataru WATANABE, Jumpei ANDO, Keisuke KAWAUCHI, Toshiyuki ONO
Life of this patent
Estimating the term…
Abstract
According to one embodiment, a data analysis apparatus includes processing circuitry. The processing circuitry acquires a data element group composed of a plurality of data elements, calculates, in regard to a first data element in the data element group, an index value indicative of a conditional dispersion from another data element in the data element group, extracts, based on the index value, a parent element having a high association with the first data element from the data element group, and outputs an analysis result including first information relating to the first data element and the parent element.
Claims18
A data analysis apparatus comprising processing circuitry configured to:
- acquire a data element group composed of a plurality of data elements;
- calculate, in regard to a first data element in the data element group, an index value indicative of a conditional dispersion from another data element in the data element group;
- extract, based on the index value, a parent element having a high association with the first data element from the data element group;
- and
- output an analysis result including first information relating to the first data element and the parent element.
The data analysis apparatus according to claim wherein
- when a plurality of parent elements are extracted, the processing circuitry is further configured to:
- calculate the index value indicative of a conditional dispersion between elements in regard to combinations of the parent elements;
- and
- extract a parent-child relationship between the parent elements, based on the index value, and
- the analysis result includes a hierarchical structure based on the first data element and the parent elements.
A data analysis apparatus comprising processing circuitry configured to:
- acquire a data element group that is composed of a plurality of data elements and that relates to a change of a specific data element, and information of a parent element having a high association with each data element of the data element group;
- estimate a causal element that is a cause of the change, by executing regression analysis by using the specific data element and the data element group;
- extract, based on the information of the parent element, a parent element having a high association with the causal element from the data element group;
- calculate an effect size on the change of the specific data element, in regard to the causal element and the parent element corresponding to the causal element;
- and
- output an analysis result including first information relating to the causal element and the corresponding parent element, and second information relating to the effect size.
A data analysis method comprising:
- acquiring a data element group composed of a plurality of data elements;
- calculating, in regard to a first data element in the data element group, an index value indicative of a conditional dispersion from another data element in the data element group;
- extracting, based on the index value, a parent element having a high association with the first data element from the data element group;
- and
- outputting an analysis result including first information relating to the first data element and the parent element.
Dependencies are inferred from the claim text.