Automatic solution model of mathematical word problems based on DAGCN and R-Drop

With the rapid development of artificial intelligence and intelligent algorithms, intelligent technology is being used more a nd more widely in various fields. Traditional mathematical word problem solving models cannot meet the needs of automatic processing of complex problems. For this reason, this study designed a new mathematical word problem solving model based on a deep adaptive graph convolutional network and an improved Dropout algorithm to solve the problems of poor solving effect and low solving accuracy of traditional mathematical word problems. The new model uses a deep adaptive graph convolutional network to improve the model's solving and computing capabilities for mathematical word problems, and uses an improved Dropout algorithm to improve the training effect of the solving model. The research results show that the solving accuracy of the research model is improved by 12.6% and 12.8% respectively compared with the sequence -to-sequence model, and the research model has better stability. It can be seen that the research model can effectively improve the solving accur acy of the traditional mathematical word problem automatic solving model, which has a good guiding significance for the automatic solving of mathematical word problems.