Scholarlink

<p>Following the COVID-19 outbreak, the education system has shifted from traditional classrooms to virtual learning. With the rapid rise of online learning platforms, the main challenge lies in user disengagement. Artificial intelligence and adaptive education systems help to personalize learning experiences and bridge educational gaps. In this paper, we present a real-time text prediction system developed using particle swarm optimization (PSO) and long short-term memory (LSTM) on an intelligent search engine from the Open Classroom (OCR) initiative. The proposed system uses the PSO method to optimize the hyperparameters of an LSTM. We demonstrate that the PSO-optimized LSTM architecture supports real-time text predictions, allowing the system to predict the next word during searches based on user input. This significantly improves user experience and engagement. Compared to manually tuned models, our model achieved substantially better performance, with a prediction accuracy of 95%. Both the e-learning system and the OCR initiative demonstrated notable impacts. The system’s accuracy and effectiveness were verified using precision and recall metrics, demonstrating its strong prediction capabilities. Additionally, it displayed lower computational latency, an asset for real-time use. These results, supported by extensive testing over multiple periods, demonstrate that the PSO-optimized LSTM is a valuable tool for innovative learning environments and opens new avenues for research and improvement.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Advanced Text Prediction System Integrated Within the Search Engine for the Open Classroom Approach Based on Particle Swarm Optimization and Long Short-Term Memory Models

  • 期刊论文

  • 发表日期：

    2026年3月26日

查看全文
  • Ines Boutabia,
  • Abdelmadjid Benmachiche,
  • Ali Abdelatif Betouil,
  • Chaouki Chemam,
  • Khadija Rais

摘要

Following the COVID-19 outbreak, the education system has shifted from traditional classrooms to virtual learning. With the rapid rise of online learning platforms, the main challenge lies in user disengagement. Artificial intelligence and adaptive education systems help to personalize learning experiences and bridge educational gaps. In this paper, we present a real-time text prediction system developed using particle swarm optimization (PSO) and long short-term memory (LSTM) on an intelligent search engine from the Open Classroom (OCR) initiative. The proposed system uses the PSO method to optimize the hyperparameters of an LSTM. We demonstrate that the PSO-optimized LSTM architecture supports real-time text predictions, allowing the system to predict the next word during searches based on user input. This significantly improves user experience and engagement. Compared to manually tuned models, our model achieved substantially better performance, with a prediction accuracy of 95%. Both the e-learning system and the OCR initiative demonstrated notable impacts. The system’s accuracy and effectiveness were verified using precision and recall metrics, demonstrating its strong prediction capabilities. Additionally, it displayed lower computational latency, an asset for real-time use. These results, supported by extensive testing over multiple periods, demonstrate that the PSO-optimized LSTM is a valuable tool for innovative learning environments and opens new avenues for research and improvement.