Vocational education skill assessment and intelligent assistance involve evaluating people' talent in precise vocational abilities and conveying personalized assist to enhance studying results. The need for such assessment and assistance arises from the significance of appropriately evaluating learners’ readiness and proficiency in vocational abilities, identifying areas for improvement in teaching practices, and presenting timely feedback and guidance to learners. However, present strategies often depend on conventional assessment techniques which can lack granularity and fail to provide personalised assistance. To address those demanding situations, this study introduces a novel method that integrates SMOTE data processing, Federated LSTM (Fed-LSTM) for skill word extraction and classification, and fuzzy rule-based vocational education talent evaluation. This approach targets to overcome class imbalances in datasets through SMOTE, permit collaborative learning across distributed data sources, and improve the accuracy and robustness of talent assessment models. The proposed study improves data representation, facilitating collaborative learning, enhancing skill extraction accuracy, and presenting robust skill assessment. The results of study are applied in a Python software, offering educators and stakeholders a realistic approach to enhance vocational education skill assessment and intelligent assistance. The proposed Fed-LSTM technique demonstrates a substantial growth in accuracy compared to the LSTM approach. With an accuracy of 99.4%, the proposed technique considerably outperforms the LSTM method, which achieved an accuracy of 76. 98%. This represents a substantial improvement of 22.42% in accuracy. © 2024 Little Lion Scientific. All rights reserved.
Author Keywords
Index Keywords
