Abstract:
Relation extraction is the process of scanning text for relationships between named entities. Recently, significant studies have focused on automatically extracting relations from biomedical corpora. Most existing biomedical relation extractors require manual creation of biomedical lexicons or parsing templates based on domain knowledge. In this study, we propose to use kernel-based learning methods to automatically extract biomedical relations from literature text. We develop a framework of kernel-based learning for biomedical relation extraction. In particular, we modified the standard tree kernel function by incorporating a trace kernel to capture richer contextual information. In our experiments on a biomedical corpus, we compare different kernel functions for biomedical relation detection and classification. The experimental results show that a tree kernel outperforms word and sequence kernels for relation detection, our trace-tree kernel outperforms the standard tree kernel, and a composite kernel outperforms individual kernels for relation extraction.
Abstract:
Most information retrieval techniques have been developed for English and other Western languages. As the second largest Internet language, Chinese provides a good setting for study of how search engine techniques developed for English could be generalized for use in other languages to facilitate Internet searching and browsing in a multilingual world. This paper reviews different techniques used in search engines and proposes an integrated approach to development of a Chinese medical portal: CMedPort. The techniques integrated into CMedPort include meta-search engines, cross-regional search, summarization and categorization. A user study was conducted to compare the effectiveness, efficiency and user satisfaction of CMedPort and three major Chinese search engines. Preliminary results from the user study show that CMedPort achieves similar accuracy in searching tasks, and higher effectiveness and efficiency in browsing tasks than Openfind, a Taiwan search engine portal. We believe that the proposed approach can be used to support Chinese information seeking in Web-based digital library applications. © Springer-Verlag Berlin Heidelberg 2003.
Abstract:
The Kohonen Self-Organizing Map (SOM) is an unsupervised learning technique for summarizing high-dimensional data. When applied to textual data, SOM has been shown to be able to group together related concepts in a data collection. This article presents research in which we sought to validate this property of SOM, called the Proximity Hypothesis. We demonstrated that the Kohonen SOM was able to perform concept clustering effectively, based on its concept precision and recall scores judged by human experts. We believe this research has established the Kohonen SOM algorithm a promising textual classification technique for addressing the long-standing `information overload' problem.
Abstract:
A proposed lexicon-enhanced method for sentiment classification combines machine-learning and semantic-orientation approaches into one framework that significantly improves sentiment-classification performance. © 2010 IEEE.
Abstract:
An overview of several digital government projects and initiatives that combine the technical and conceptual threads was presented. The project aimed to make Federal statistical data more easily available and usable by the broadest possible audiences. A framework for mapping questions onto interface mechanisms that depended on using metadata as an intermediary between user needs and agency data was also developed.
