Ontology based news extraction system using recurrent neural networks

24-hour news is the new trend among news channels which gets updated every second. Enormous amounts of news refresh set the challenge of archiving it. Audience are enthusiastic about news stories for its crisp yet detailing quality, the focus of these stories gives a kaleidoscopic view from the past to the latest update on the topic. News archives serves to function as a pool of primary details for further modifications and updates. Eventhough the news correspondent gets the semantically related news for search query, extracting and linking related information is a tedious and time consuming job and remains a major drawback of the system. Thus we present a personalized news extraction framework based onontology and sequence to sequence learning with neural networks that automatically creates news for a query. The systemis tested using the metrics BLEU and ROUGE. Finally, the metrics score is compared with human evaluation.