Grapharizer: A Graph-Based Technique for Extractive Multi-Document Summarization

In the age of big data, there is increasing growth data on Internet. It becomes frustrating for users to locate desired data. Therefore, text summarization emerges as a solution this problem. summarizes and presents with gist provided documents. However, summarizer systems face challenges, such poor grammaticality, missing important information, redundancy, particularly in multi-document summarization. This study involves development graph-based extractive generic MDS technique, named Grapharizer (GRAPH-based summARIZER), focusing resolving these challenges. addresses grammaticality problems summary using lemmatization during pre-processing. Furthermore, synonym mapping, multi-word expression anaphora cataphora resolution, contribute positively improving generated summary. Challenges, redundancy proper coverage all topics, are dealt achieve informativity representativeness. novel approach which can also be used combination different machine learning models. The system was tested DUC 2004 Recent News Article datasets against various state-of-the-art techniques. Use increased accuracy by up 23.05% compared baseline techniques ROUGE scores. Expert evaluation proposed indicated more than 55%.