Data analyzing device
To make it easier to perform an in-depth analysis by presenting to a user which feature value influences and how it influences an objective variable in a case of automatically generating feature values. For each of a plurality of feature values, by determining a division point indicating a change in influence on an objective variable, each feature value is divided into a plurality of segments and an evaluation value is calculated using an influence degree of each segment on the objective variable as an index. The segments of the feature values for which the evaluation values have been calculated are displayed on a monitor.
1. A data analyzing device for analyzing analysis target data, comprising:
a data input unit which receives an input of analysis target data including a plurality of attributes,
a feature value automatic generating unit which automatically generates a new feature value by applying a predetermined function to a single attribute of the plurality of attributes or a combination of the plurality of attributes,
an analyzing unit which generates a prediction model for predicting an objective variable based on a plurality of feature values including a feature value of an original attribute included in the analysis target data and a new feature value generated by the feature value automatic generating unit,
an evaluating unit which, for each of the plurality of feature values, by determining a division point indicating a change in influence on the objective variable, divides each of the plurality of feature values into a plurality of segments and calculates an evaluation value using an influence degree of each segment of the plurality of segments on the objective variable as an index, and
a display unit which displays the plurality of segments of the plurality of feature values for which the evaluation value has been calculated by the evaluating unit.
2. The data analyzing device according to claim 1 , wherein the display unit is configured to selectively display the plurality of segments of the plurality of feature values for which the evaluation value has been calculated by the evaluating unit.
3. The data analyzing device according to claim 1 , wherein the display unit is configured to preferentially display the plurality of segments of the plurality of feature values for which the evaluation value has been calculated by the evaluating unit.
4. The data analyzing device according to claim 1 , wherein the display unit is configured to display a segment of the plurality of segments of a feature value of the plurality of feature values having a high-ranking evaluation value calculated by the evaluating unit.
5. The data analyzing device according to claim 1 , wherein the evaluating unit is configured to determine the division point where the evaluation value is optimal.
6. The data analyzing device according to claim 1 ,
wherein the evaluating unit is configured to, by determining the division point indicating the change in influence on the objective variable under a compound condition combining the plurality of feature values, divide each of the plurality of feature values into the plurality of segments and calculate the evaluation value using the influence degree of each segment of the plurality of segments on the objective variable as the index, and
the display unit is configured to display the plurality of segments of the plurality of feature values for which the evaluation value under the compound condition has been calculated.
7. The data analyzing device according to claim 1 ,
wherein the evaluating unit is configured to calculate a share of data included in the plurality of segments, and
the display unit is configured to display the share calculated by the evaluating unit.
8. The data analyzing device according to claim 1 ,
wherein the evaluating unit is configured to calculate an average value of objective variables, and
the display unit is configured to display the average value of the objective variables.
9. The data analyzing device according to claim 1 , wherein the display unit is configured to be capable of displaying switching between a first display form in which a segment of the plurality of segments having one feature value is displayed and a second display form in which another segment of the plurality of segments having a plurality of feature values is displayed.
10. The data analyzing device according to claim 1 ,
wherein the display unit is configured to display a relation between a representative value of the plurality of feature values used in the prediction model and a numeric value of the objective variable, and
the data analyzing device further comprises:
an adjusting unit which receives an adjustment of the representative value based on a user's operation input, and
a simulation unit which changes the plurality of feature values according to an adjustment amount of the representative value, recalculates the numeric value of the objective variable from the prediction model and displays on the display unit.