Effective Integration of Automatic Word Spacing and Morphological Analysis in Korean

In general, morphological analysis in Korean is performed on a sentence in which word spaces are correctly inserted. To process a sentence that is not correctly spaced, automatic word spacing model should be performed in advance. many previous studies have adopted a pipeline architecture in which results of word spacing are used as inputs of morphological analysis. However, under this kind of pipeline architecture, word spacing errors lead to dismissed performances of morphological analysis models. To resolve this problem, we propose an integrated neural network model that performs word spacing and morphological analysis at the same time. The proposed model consists of two layers of bidirectional gated recurrent unit models with conditional random field layers: a lower layer for word spacing and an upper layer for morphological analysis. In the experiments, the proposed model outperformed both an independent word spacing model and independent morphological analysis model. It also showed better performances than the previous integrated models.