Predicting Efl Learners' Writing Proficiency Through Reading and Writing Practices: An Artificial Neural Networks Approach

Abstract The current quantitative study explores the extent to which reading and writing practices predict the writing proficiency of 7th-grade EFL learners through Artificial Neural Networks in order to investigate nonlinear interactions among reading comprehension, writing volume, and writing proficiency. The participants of the study consisted of 173 students, and predictive efficacy was assessed using Multiple Linear Regression (MLR), Artificial Neural Network (ANN) and Radial-Basis Neural Networks (RBNNs) models. The obtained data were analysed using skewness and kurtosis values, as well as Kolmogorov-Smirnov and Shapiro-Wilk tests. The results indicate that reading comprehension and writing volume are strong predictors of writing proficiency, and ANN models perform better than traditional regression in capturing complex interdependencies. These findings have important pedagogical implications for EFL curriculum design and recommend systematic reading and writing interventions along with machine learning assessments and personalized instruction. Keywords: Writing proficiency, artificial neural networks, machine learning, EFL learners