Assessing English language sentences readability using machine learning models.

Area of Science:

• Natural Language Processing
• Machine Learning
• Computational Linguistics

Background:

• Text readability assessment is crucial for effective communication and education.
• Machine learning offers advanced capabilities for text analysis and classification.
• Previous readability research has laid the groundwork for data-driven approaches.

Purpose of the Study:

• To develop and evaluate machine learning models for English text readability assessment.
• To predict the literacy level required for understanding given sentences.
• To enhance English language learning by providing suitable reading materials.

Main Methods:

• Utilized a dataset of 30,000 English sentences.
• Annotated sentences into seven readability levels using the Flesch Kincaid method.
• Experimented with five machine learning algorithms: KNN, SVM, LR, NB, and ANN.

Main Results:

• All tested classification models demonstrated excellent and stable performance.
• The Artificial Neural Network (ANN) model achieved the highest F-score of 0.95% on the test set.
• The models effectively classified text readability across different levels.

Conclusions:

• Machine learning models, particularly ANN, are highly effective for readability assessment.
• The developed model can be integrated into educational settings to support language learning.
• Accurate readability prediction aids in tailoring content for English Language Learners and assessing their abilities.