Automatic Abstraction Rating of Research Papers using Hierarchical Convolutional Neural Network

Abstract

In every there are many papers submitted to publishers. In that time, reviewer review that papers by reading and then give an output like accepted or rejected. Reviewers read every paper one by one and give a review about that paper is a time consuming. This research only worked for reviewing the abstraction of a paper. If we review an abstraction by using a model it will be very effective for saving time. In this research work, we discuss classification algorithms from machine learning algorithms for finding better result. We make a data set for our model from different review system. Many kinds of models or algorithms are used for this work. The most important usable algorithms are modularized Hierarchical Convolutional neural network and another is an attention-based convolutional neural network. We combine this two algorithms and named as HCNN for reviewing abstract of papers.