Related Experiment Video
Updated: Sep 29, 2025
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
665
English Text Readability Measurement Based on Convolutional Neural Network: A Hybrid Network Model
Lihua Jian1, Huiqun Xiang2,3, Guobin Le2,3
1School of International Education, Hunan University of Medicine, Hunan, Huaihua 418000, China.
Computational Intelligence and Neuroscience
|March 25, 2022
Summary
This study introduces a hybrid deep learning model for measuring English text readability, improving efficiency and performance over traditional manual methods. The model combines convolutional neural networks, bidirectional long short-term memory networks, and attention mechanisms for automated feature extraction.
Area of Science:
- Natural Language Processing
- Computational Linguistics
- Artificial Intelligence
Background:
- Text readability is crucial for information accessibility, with increasing demand for accurate measurement tools.
- Traditional readability assessment relies on manual feature extraction, which is time-consuming, subjective, and prone to performance degradation due to feature redundancy.
- The growing volume and complexity of digital text necessitate advanced, automated methods for readability analysis.
Purpose of the Study:
- To propose a novel hybrid deep learning model for automated English text readability measurement.
- To overcome the limitations of manual feature engineering in traditional readability assessment.
- To enhance the efficiency and accuracy of text readability evaluation.
Main Methods:
- Development of a hybrid deep learning network integrating Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM) networks, and Attention Mechanisms.
- Application of the hybrid model to analyze text structure at word, sentence, and document levels.
- Automated feature extraction using machine learning, replacing manual expert-driven feature engineering.
Main Results:
- The proposed hybrid network model significantly improves the efficiency of English text readability measurement.
- Automated feature extraction via the deep learning model enhances overall measurement performance compared to traditional methods.
- The model effectively captures deep features from text structure, leading to more accurate readability scores.
Conclusions:
- Hybrid deep learning models offer a superior approach to English text readability measurement.
- Automated feature extraction is key to improving the scalability and performance of readability assessment tools.
- The developed CNN-BiLSTM-Attention model provides an effective solution for the growing demand for text readability analysis.
