Research On Automatic Solving Technology Of Mathematical Word Problems Integrating Seq2Tre

With the continuous development and popularization of artificial intelligence technology,intelligent education has gradually become an important trend in the field of education.In this trend,mathematical problem solving has become a research focus,aiming to integrate human mathematical knowledge into the problem-solving process to improve the efficiency and quality of mathematical learning.However,a challenge of mathematical problem solving is how to effectively represent and apply human mathematical knowledge.Traditional rule-based methods require manual rule writing,while deep learning methods can learn mathematical knowledge from large amounts of data,but their generated solutions lack interpretability and are difficult to generate effective equations for multi-operator problems.In mathematics education,traditional question banks are a major source of exercises and test questions for students to practice and test their knowledge.However,traditional question banks have some drawbacks,such as static questions and being out of touch with the latest teaching methods.In response to these issues,question bank websites have emerged.These websites are usually online platforms that provide a wide range of problems and assessments for various subjects,including mathematics and science.Question bank platforms help to overcome some of the limitations of traditional question banks by providing students with more problems and more detailed evaluations.However,question bank platforms may still have limitations in providing indepth explanations and supporting individual learning needs.Traditional question bank platforms have problems such as insufficient personalized recommendations and explanations,low levels of automation and intelligence,which cannot provide students with accurate and efficient learning experiences and assistance.To address these issues,this article introduces the Seq2 Tree model for automatic solving of mathematical application problems,with the following two main tasks:(1)To enhance the interpretability of automated problem solving and improve the accuracy of equation prediction for multi-operator problems,this paper proposes a formula-enhanced Seq2 Tree model-the FEST model.The FEST model design incorporates formulas into the expression tree node generation,allowing expression tree nodes to be preset with formulas to generate an indefinite number of nodes,optimizing the predicted equation length and reducing equation complexity.In the encoding and decoding process,information is supplemented using sibling nodes,and a stack is used to monitor the progress of prefix expression generation and automatically terminate the decoding process when the tree is completed,improving the accuracy of generating effective equation templates.To verify whether formulas are helpful for deep learning networks to automatically answer mathematical application problems,this paper extracted several geometry problems that use formulas most frequently from the Math23 K dataset to form a new dataset GMath23 K and annotated it with formulas.The experimental results show that for problems that require mathematical formula knowledge,the FEST model can provide better performance and more interpretable results.(2)An automated problem-solving platform can automatically recommend suitable problems and problem-solving methods based on students’ personalized needs and learning situations,and provide more detailed problem-solving processes and explanations to help students understand and master mathematical knowledge.To further improve the automation and interpretability of mathematical problem solving,this paper developed a mathematical application problem automatic solving platform based on the FEST model.This platform mainly targets mathematical application problems,uses Vue,Flask,and Mongo DB database technology,combines the formulaic problem-solving steps of the FEST model,and provides automatic problem solving and high interpretability problem-solving ideas to help students learn and master knowledge,and improve learning outcomes and problem-solving abilities.