차원축소를 적용한 효율적인 개체명 인식을 위한 Word Embedding 방법 연구/A Study on Word Embedding Method for Efficient Named Entity Recognition with Dimension Reduction

In name entity recognition, the Word Embedding dimension size, which is a hyper-parameter, affects learning time and performance. But learning time and performance compete with each other, it is difficult to reduce learning time while maintaining performance. In this paper, we propose a method to reduce the size of Word Embedding using PCA and Kernel PCA. It also demonstrates how to enter and train more diverse qualities than the existing Word Embedding method.