The Comparison Of Pooling Functions In Convolutional Neural Network For Sentiment Analysis Task

Convolutional Neural Network (CNN) has gained considerable attention in many Natural Language Processing applications including sentiment analysis task. A typical CNN model usually is made up of several convolutional and pooling layers. In this paper, our aim is to acquire detailed understanding into different type of pooling function by directly differentiate them on a same architecture layers for sentiment analysis tasks. These insights should prove useful for future development of pooling function in CNN models for sentiment analysis task.