Decision Tree for Predicting Student English Writing Performance in the Higher Education

Decision tree is a machine learning algorithm that can effectively predict student performance. However, the existing performance prediction models rarely analyze the impact of multiple factors on English writing performance. To address this issue, we propose an English writing performance prediction model based on the CART decision tree algorithm. Specifically, we first collect data on motivation, engagement, self-efficacy, parent education, and years of learning English from 974 English major students in four universities in Yunnan Province, China. Then, we preprocess the data by cleaning, transforming, and splitting it into training and testing sets. Next, we build a CART decision tree model using Python 3.0 and the scikit-learn and pandas libraries, and compare it with support vector machines and logistic regression to evaluate its accuracy. Finally, we analyze the results to identify the important factors affecting students' English writing performance. The experimental results show that the years of learning English, engagement, and parent education are the key factors in predicting writing performance. Overall, our study enriches the existing literature by clarifying the complex factors that influence students' English writing performance and provides an exemplary application of decision tree-based quantitative methods in the realm of English education research.