Development of an Enhanced Naïve Bayes and Association Rule Mining Hybridized Algorithms for Multi-Documents Summarization Approach

The challenge of summarization arises from the fact that computers lack the ability to comprehend human language and the emotions conveyed within it. To address this challenge, several machine learning models have been trained, tested and used in a manner that encompasses crucial information. Researchers were drawn to various aspects of artificial text summarization due to its ability to efficiently extract relevant information in a shorter timeframe and its uses encompasses the generation of concise summaries from many sources such as emails, news articles, mobile news, and corporate information. One hundred thousand (100,000) records of case related to heart disease were fetched from Kaggle online Dataset for Heart Disease, the data were contained in different types of documents They were fed into Naïve Bayes algorithm for proper data cleansing and treatment before classification into 70% training and 30% testing because of the status of the model. Association rule mining was used to collaborate to achieve very reliable summaries. The enhance Naive Bayes classifier was used as initiator to provide set of relevance scores by analyzing word frequencies and other document features. These scores were then used as part of the broader summarization framework, the two output were combined to generate a comprehensive and meaningful summary. Performance evaluation were carried out. The performance evaluation metrics adopted in the study measures the model’s accuracy, precision, recall and F1 score, the model, combines both enhanced Naive Bayes (e-NB) and Association Rule Mining (ARM), delivering superior performance across all the metrics. Overall performance of (e-NB) - ARM outshines NB-ARM in all metrics, including accuracy, precision, recall, F1-score, and time complexity, the hybridized model improves predictive accuracy by combining the probabilistic approach of Naive Bayes with the insightful patterns generated by Association Rule Mining, allowing for more robust predictions, accurate and interpretable insights of heart disease making it the preferable choice between the two models. Therefore, adopting the enhanced hybrid Naive Bayes and Association Rule Mining (e-NB) - ARM system for multi-document summarization is the best. Conclusively, this approach not only enhances the efficiency and quality of multi-document summarization in high-volume textual data environments but also has applications in diverse fields. This study has contributed to the growing need for advanced text summarization methods, facilitating faster access to essential information and supporting more informed decision-making across disciplines.