A Hybrid Text Summarization Approach Using Neural Networks and Metaheuristic Algorithms.

The rapid growth of text data on the Internet requires effective automatic text summarization techniques. This study proposes a hybrid text summarization approach that combines a Multi-hidden Recurrent Neural Network and a mayfly-harmony search algorithm. The neural network generates a feature vector for each sentence. The mayflyharmony search algorithm then optimizes the feature weights to extract the most relevant sentences for the summary. This manuscript capacity provides essential information and expertise that can be effectively summarised using Efficient Abstractive Text Summarising (EATS) techniques. This project aimed to extract informative summaries from various articles by utilizing regularly utilized handcrafted elements from literature. A Multi-hidden Recurrent Neural Network (MRNN) was used to generate a feature vector, and a new feature assortment strategy called Mayfly-Harmony Search (MHS) was applied for feature extraction. The number of sentences, word frequency, title similarity, term frequency-inverse sentence frequency, sentence location, sentence length, sentencesentence similarity, sentence phrases, proper nouns, n-gram co-occurrence, and document length were the features used. By taking diverse Mayfly Algorithm explanations found from other expanses of the search space and processing them with Harmony Search, the suggested hybrid of the Mayfly Algorithm and Harmony Search was employed to produce superior results. [ABSTRACT FROM AUTHOR]