Keyphrase extraction by the use of glove and ResNeXt optimized by enhanced human evolutionary optimization (EHEO)...

Area of Science:

• Natural Language Processing
• Artificial Intelligence
• Machine Learning

Background:

• Keyphrase extraction (KPE) is crucial for document summarization, search engine optimization, and information retrieval.
• Manual KPE is time-consuming, and automated methods struggle with contextual understanding.
• Existing automated KPE techniques require improvement in accuracy and efficiency.

Purpose of the Study:

• To develop an innovative and effective automated keyphrase extraction method.
• To enhance the understanding of contextual relationships in text for KPE.
• To provide a robust solution for automatic keyphrase extraction applicable to various fields.

Main Methods:

• Employed the ResNeXt neural network architecture.
• Optimized the model using an enhanced human evolutionary optimization algorithm.
• Integrated GloVe-100 word embeddings for contextual representation.
• Evaluated the model on KP20k, Inspec, and SemEval-2010 datasets.

Main Results:

• Achieved superior performance compared to advanced methods like BERT and CNN-BERT.
• Attained recall, precision, and F1-score values of 98.81%, 98.67%, and 98.74% on the KP20k dataset.
• Demonstrated high performance on Inspec (96.54% precision, 96.32% recall, 96.43% F1-score) and SemEval-2010 (97.32% precision, 97.81% recall, 97.56% F1-score).

Conclusions:

• The integration of high-quality embeddings and optimization strategies with advanced neural architectures is effective for KPE.
• The proposed model offers a strong and efficient solution for automatic keyphrase extraction.
• The method shows potential for broad application in diverse fields requiring text analysis.