ABSTRACT

Due to switch on online mode the study of summarization of various domain has grown during the COVID-19 pandemic, where education is no exception. In order to open an alternative learning opportunities for the students and support their parents, we have understood the importance of summarization system. Hence, we are motivated to design and automated summarization system which summarized the conversations between the speakers. There are many exchanges between speakers in conversations, which are typically referred to as dialogue acts. The reader will be frustrated and bored with dialogue that functions as a data dump or contains boring life details. Using simplified conversation is one method to avoid needless dialogue. In this work, we have developed an automated dialogue summarization system in recent trends COVID-19, where we have selected three types of conversations such as online classes, social distancing, and career opportunities. In order to build this system, firstly we have converted the input dialogues into text depending on end break points or punctuations. Thereafter, we have employed a well-known text rank algorithm to identify the important sentences from the dialogues. Finally, the extracted sentences are arranged to prepare the summarized output which is understandable and meaningful.

IndexTerms—Abstractive Summarization, Extractive Summa-rization, Glove, Rouge score