Apparatus and method for learning data construction

The present invention for providing an information search, information retrieval, translation, training data to build apparatus and method for constructing the training data required by the statistical methodology for the operations, such as natural language processing more efficiently, (a) the machine with respect to the training data a step of performing learning generates a learning model and, (b) and the step of generating the training data candidate to automatically attach the tag to the source corpus by using the learning models of the generation, (c) the created learning data candidate and the step of calculating a confidence score, and using the confidence score of the calculated candidate select learning data candidate, (d) correct errors in the selected learning data candidates via the user interface, and wherein the error corrected the study Add the candidate data in the learning data can includes the step of gradually expanding to new learning models . Learning data, automatic tagging, data, Learning candidate selection, active learning, learning progress