A computer implemented method for a relabeling method for regression analysis, comprising: (a) constructing an initial regression analysis model to which a preselected regression analysis method is applied to a predetermined data set, wherein The data set comprises a set of data given a result value and data without a result value; (b) selecting data which is an instance of a label object among data sets of the constructed regression analysis model; (c) measuring the resultant value of the selected data according to the regression analysis model constructed, and if the selected data is data that does not already exist, label the resultant value anew and the selected data is Labeling a new value with an average value between the existing result value and the newly measured result value when the result value data exists previously; (d) constructing a new data set that updates the newly labeled data in the existing data set, and reconstructing the regression analysis model using the constructed new data set; And (e) repeating steps (b) to (d) until the condition for specifying a regression analysis end condition is satisfied, thereby constructing a final regression analysis model. .
展开▼