Abstract:
Text Summarization is a strategy of summarizing any passage, document or text automatically. Summarized text is nothing but the minimized form of the given text. There are many techniques available for the English text summarization but for Bangla there are few works exits. Our main purpose of this project is to generate an understandable and meaningful summary which is fluent and easy for the people. Because language is the main obstacle for communication. Text summarization can help reduce the time it takes to read and understand long texts by providing a condensed version of the content. Summarizing text can help clarify the main points and improve overall comprehension of the material. We have collected data from kaggle, a web platform and newspaper. To get an outline we've got to use our model. Our model is Sequence-to-Sequence supported bi-directional RNN with LSTM. Throughout this project we've got some problems like preprocessing, vocabulary count, missing words count, word embedding and so on. During this project, our main goal is to deduce the operating loss and build a fluent outline and build a higher method for Bangla text summarization. We tend to area units able to cut back the loss worth below 0.031. Our model accuracy is 97.69%. Our model is ready to make theoretical text summarization.
