34 The Modification of the LRV Algorithm in ESA for Alleviating Data Sparseness
Download citation file:
- Ris (Zotero)
- Reference Manager
Word Segmentation is the very first stage for processing some languages, such as Chinese. We early proposed an unsupervised approach to word segmentation. In this article, we analyze a part of the core algorithm of the approach to demonstrate how the part alleviates the sparse data problem. By experimental results, the data sparseness really does negative effect on the performance of the approach. Meanwhile, the results also indicate that the part of the algorithm has the capacity to reduce the influence of data sparseness.