A Hybrid Approach for Automatic Text Summarization and Translation based On Luhn, Pegasus, and Textrank Algorithms

Text summarization equipment is in high demand these days due to the fact the primary goal of a textual content summarization gadget is to become aware of the most crucial facts from a given textual content and gift it to stop customers. In this project, we deployed a web application that accepts any generic paragraph as input to provide a summary version of that particular paragraph by identifying textual features that are of high significance and also translating the generated summarized text into any language. To summarize the text, we proposed a hybrid model that is based on the Luhn and Textrank algorithms, which are extractive summarization techniques, and the Pegasus model, which is an abstractive summarization technique. This hybrid model was also compared with BERT, XLNet, and GPT2 based on their ROGUE scores. The generated optimal paragraph is sent as input to the translator, which translates the summarized text into any language. In this project, we also tried to get better results from the proposed hybrid model by comparing it with other existing models. Firstly, the TextRank algorithm is used for arranging the sentences based on their priority, and then abstractive summarization is performed on this paragraph using the Pegasus model to generate a new summary with good context, which is then fed to the Luhn algorithm, which generates the final optimal paragraph. The proposed hybrid model achieved higher ROGUE scores in comparison with other existing models such as BERT, GPT2, and XLNet.