Deep Learning Algorithm using Text Classification

Abstract

Text classification based on nature language process (NLP) is considered the process of labeling a textual document with most applicable tags. However, the amount of texting in social media magazines, newspapers, blogs, and websites are becoming a huge difficulty to be analyzed. In this case, it is important to find a proper algorithm that can provide an optimal accuracy and performance. Neural networks and deep learning are considered popular algorithms at the current era. These methods have already solved many problems that seemed to be too hard. This study attempts to find the best algorithms based on text classification. This paper used text classification based on CNN (convolutional neural network), RNN (Recurrent Neural Network) & HAN (Hierarchical Attention Network) algorithms. In this paper, 100 datasets are collected (80 for training and for 20 testing). The CNN algorithm obtained the best accuracy and performance with lowest in regards with loss, these were indicated as results.