Classification of Texts Using LSTM and LDA

Abstract

With the approach of the Fourth Industrial Revolution, information has become a powerful resource for society and the economy to operate and develop. In particular, as customized algorithms have become an essential service in most systems, the importance of big data-based information processing technology is also deepening. However, human languages have extreme variability compared to programming languages, so interpretation and processing are difficult. Efficient measures need to be taken to extract the desired information from these unstructured data. Therefore, in this paper, we identify and develop a more effective analytical system by classifying papers that are subdivided into five categories within the topic of 'technique' into LSTM and LDA techniques after learning.