Mathematical problem concept type prediction service provision method using neural network-based machine translation and mass corpus

摘要

To provide a method for providing prediction service of a mathematical problem concept type using neural machine translation and math corpus.SOLUTION: A system for providing prediction service of a mathematical problem concept type comprises the steps of: receiving a mathematical word problem; separating the mathematical word problem into a text and an image by using natural language processing and image processing; analyzing the separated text by using morpheme analysis and individual name recognition on the basis of math corpus, analyzing the image by using an object recognition and semantic analysis, and performing mathematization translation to extract a math term; filtering, extracting, and compressing a concept type candidate group on the basis of the mathematization-translated math term; analyzing the mathematization-translated math term to be classified into preset cases and terms; and classifying the concept type of the mathematical word problem by analyzing a syntax pattern on the basis of preset calculation problem pattern, a word problem pattern, and explanatory problem pattern definition by using neural network-based machine translation.SELECTED DRAWING: Figure 7