Ensembled Approach for Text Summarization

There is a tremendous amount of data which is present online, and to extract the useful content is a challenging task. The solution is made possible by the introduction of text summarization. In this paper, an ensembled approach for text summarization is proposed, in which the robustness of extractive summarization and abstractive summarization is combined to make the most sense out of the raw data. Extractive text summarization is implemented by using RNN model based on LSTM architecture. The output generated by this model is used as input for the abstractive summarization. Pointer Generator Network is used for implementation of abstractive text summarization. The standard CNN/daily mail dataset is used for experimental purpose. The results are evaluated using ROUGE scores.