A method performed in a system for supporting learning, the system comprising an assessment item information acquisition unit, a weak learning element determination unit, and a supplementary learning path determination unit, and the method comprising the steps of:
- by the assessment item information acquisition unit, acquiring a learner's score for at least one assessment item included in a scoring table, wherein the learner has solved a question for learning and the scoring table is applied to a descriptive answer of the learner, and the scoring table includes information on a score that a scorer gives for the at least one assessment item;
- by the weak learning element determination unit, determining a weak learning element of the learner with reference to the acquired score and at least one learning element associated with the at least one assessment item, wherein the at least one learning element is a knowledge unit that is required for the learner so that the learner is determined to fulfill or understand the at least one assessment item;
- and
- by the supplementary learning path determination unit, determining a supplementary learning path to be provided to the learner with reference to the determined weak learning element
- wherein in the step of acquiring the score, the learner's score for the at least one assessment item is acquired by acquiring a scoring result for the descriptive answer of the learner through optical character recognition (OCR),
- wherein in the step of determining the weak learning element, the at least one learning element associated with the at least one assessment item is acquired by analyzing at least one of words, sentences, and expressions of the at least one assessment item using a convolutional neural network algorithm, the at least one acquired learning element is coded as predetermined codes, tagging information including information in which the at least one acquired learning element is coded is assigned to the at least one assessment item such that the at least one assessment item is linked to one or more of the predetermined codes, and
- wherein in the step of determining the weak learning element, the weak learning element is determined to be a first learning element corresponding to an assessment item for which the acquired score is lower than a predetermined level among the at least one assessment item, and a second learning element having a relationship of prior knowledge or posterior knowledge with the first learning element,
- wherein in the step of determining the weak learning element, the weak learning element is determined with further reference to allotted points corresponding to the at least one assessment item in the scoring table,
- wherein in the step of determining the supplementary learning path, the supplementary learning path is specified by at least one question for supplementary learning associated with the weak learning element,
- wherein in the step of determining the supplementary learning path, the at least one question for supplementary learning includes at least one of a unitary question in which the weak learning element is contained one by one, an analogous question in which the weak learning element is combined with another learning element, and a question expected to be incorrectly answered in which the weak learning element is combined with a weak learning element determined from another question for learning solved by the learner, and
- wherein in the step of determining the supplementary learning path, the unitary question is specified as the question for supplementary learning when a frequency at which the weak learning element is determined as a weak learning element for at least one other question for learning solved by the learner is not less than a first level, the question expected to be incorrectly answered is specified as the question for supplementary learning when the frequency is less than a second level, and the analogous question is specified as the question for supplementary learning when the frequency is between the first and second levels.