Grammar Error Correction Method for English Composition Based on Spider Monkey Optimization with Multilayer Perceptron

To enhance the timeliness and accuracy of English Grammar Error Correction (EGEC), the error correction algorithm employs computer programming technology. Which aids to automatically identify and resolve grammar errors present in the English text written by nonnative language learners. The classification model enabling the extraction of information from English text data and constructing a consistent grammar correction method. In this research, an efficient EC method for English composition grammar based on Spider Monkey Optimization (SMO) with Multi-layer Perceptron (MLP) which is Deep Learning (DL) technique. Initially English composition corpus data collected and utilized the preprocessing and Term Frequency and Inverse Document Frequency (TF-IDF) extraction techniques are to extract the essential features to construct the grammatical feature set. Furthermore, the SMO implemented to tune the MLP neurons, it contributes to enhancing the model’s performance. The feature set serves as input data for the MLP and enabling feature classification. The English composition grammatical errors are detected based on high similarity score (HSS), and EC is achieved by locating the penalty parameters and minimizing the deviation parameters. Experimental outcomes are shown that the proposed method achieves syntax error detection (SED) when the experiment time is 10 minutes, the text recall rate of the English composition text is 93.23%. This performance exceeds than existing models like Convolutional Neural Network (CNN), Artificial Neural Network (ANN), and Capsule Neural Network (CapsNet).