Abstract In the modern era, Intelligent Tutoring Systems (ITS), Computer Based Trainings (CBT) etc. are gaining popularity rapidly in the educational sectors as well as professional sectors and an automatic math word problem solver is one of the crucial sub-fields of ITS. Solving mathematical word problems automatically is a challenging research problem in Artificial Intelligence (AI), Natural Language Processing (NLP) and Machine Learning (ML), since understanding and extracting relevant information from an unstructured text require lots of reasoning abilities. Till date, much research has been carried out in this domain, focusing on solving each type of mathematical word problem, which include solving like arithmetic word problems, algebraic word problems, geometric word problems, trigonometric word problems etc. In this work, we present an approach to solve arithmetic word problems automatically. However, it is limited to solve only single operation and single equation word problems. We used a rule-based approach in classifying word problems. We propose various rules to establish the relationships and dependencies among various key-entities to broadly classify the word problems into four categories (Change, Combine, Compare and Division-Multiplication) and their sub-categories to identify the desired operation among+, -, *, and /. Irrelevant information is also filtered out from input problem texts, based on hand-crafted rules to extract relevant quantities. Later, an equation is formed with the relevant quantities and predicted operation to generate the final answer. The work proposed here, performs well as compared to most of the similar systems reported on the standard SingleOp dataset achieving an accuracy of 93.02%.
Alan : Mühendislik
Dergi Türü : Uluslararası
Benzer Makaleler | Yazar | # |
---|
Makale | Yazar | # |
---|