<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-1017">
  <Title>Relation Extraction Using Label Propagation Based Semi-supervised Learning</Title>
  <Section position="4" start_page="133" end_page="134" type="metho">
    <SectionTitle>
4 Discussion
</SectionTitle>
    <Paragraph position="0"> In this paper,we have investigated a graph-based semi-supervised learning approach for relation extraction problem. Experimental results showed that the LP algorithm performs better than SVM and  bootstrapping. We have some findings from these results: The LP based relation extraction method can use the graph structure to smooth the labels of unlabeled examples. Therefore, the labels of unlabeled examples are determined not only by the nearby labeled examples, but also by nearby unlabeled examples. For supervised methods, e.g., SVM, very few labeled examples are not enough to reveal the structure of each class. Therefore they can not perform well, since the classification hyperplane was learned only from few labeled data and the coherent structure in unlabeled data was not explored when inferring class boundary. Hence, our LP-based semi-supervised method achieves better performance on both relation detection and classification when only few labeled data is available. Bootstrapping Currently most of works on the RDC task of ACE focused on supervised learning methods Culotta and Soresen (2004; Kambhatla (2004; Zhou et al. (2005). Table 5 lists a comparison on relation detection and classification of these methods. Zhou et al. (2005) reported the best result as 63.1%/49.5%/55.5% in Precision/Recall/F-measure on the relation subtype classification using feature based method, which outperforms tree kernel based method by Culotta and Soresen (2004). Compared with Zhou et al.'s method, the performance of LP algorithm is slightly lower. It may be due to that we used a much simpler feature set. Our work in this paper focuses on the investigation of a graph based semi-supervised learning algorithm for relation extraction. In the future, we would like to use more effective feature sets Zhou et al. (2005) or kernel based similarity measure with LP for relation extraction.</Paragraph>
  </Section>
class="xml-element"></Paper>