Dual embedding with input embedding and output embedding for better word representation

Recent <span lang="EN-US">studies in distributed vector representations for words have variety of ways to represent words. We propose a various using input embedding and output better than single model. compared the performance terms word analogy similarity with each embeddings dual which are combination those two embeddings. Performance evaluation results show that proposed outperform embedding, especially way simply adding figured out things this paper, i) not only but also has such meaning ii) combining as outperforms when we use individually.</span>