Multi-dimensional data mining of English online teaching platform based on improved decision tree

To improve the acceleration ratio and mining accuracy of data mining, a new multi-dimensional data mining method for English online teaching platforms is proposed based on improved decision tree. Information granule technology is introduced for data reconstruction, utilising neighbourhood data relationships to improve clustering accuracy. An association rule mapping structure is constructed using association matrix and difference coefficient matrix to present data set relationships, with mining factors and relative errors introduced to improve subsequent mining accuracy. The improved C4.5 decision tree algorithm is adopted, combined with principal component analysis to reduce dimensionality, and features are filtered through information gain rate to improve data mining accuracy and efficiency. Experimental results demonstrate significantly improved mining performance, with data mining acceleration ratio maintained above 0.9 and accuracy maintained above 98.54%.