Query-based Single Document Summarization using Hybrid Semantic and Graph-based Approach

摘要: In this era of digitization, a plethora information, spanning multitude themes, is available worldwide. The demand for query focused text summarization growing rapidly due to the presence irrelevant and repetitive information. Many researchers have immensely studied various unsupervised supervised techniques generic summarization. However, field query-based still demands significant research development. This paper presents novel hybrid methodology extract sentences that are relevant user-given question or query. proposed model uses algorithms first detect topics in data by generating an intuitive semantic structure. It then employs conditional similarity measure retrieve containing keywords. An enhanced graph-based approach has been used include additional query-relevant as well discard redundant information generate efficient output summary using weighted graph. performance measured two evaluation metrics, namely ROUGE-N/L scores peer-review survey. efficiency evaluated through coherence measure. DUC-2005 email dataset evaluation. results showcase improved compared other traditional methods encourage future work field.