Evaluating Bioinformatics Technology

提案：0326404PI: Lynette HirschmanTitle：评价生物信息学技术摘要这项工作的重点是建立一个文本数据挖掘的国际“挑战评估”，模仿成功的CASP评估（蛋白质结构预测技术的关键评估）。初步评价将侧重于两项任务：确定具有重要生物学意义的实体；以及提取关系信息，如文献中基因本体术语的相互作用或识别。通过使用来自BIND、SWISS-PROT和其他模式生物数据库的专家注释数据，将获得训练和测试语料库的金标准。这些数据集将通过一个网站提供给广大社区。计算语言学和信息检索领域的组织也将参与其中。总体策略是将重点放在对工作中的生物学家重要的问题上，例如改进生物数据库的管理工具，以及更好地获取文献和管理数据库中的文本信息。这种方法为多个研究小组提供了基础设施和支持。它还支持创建训练和测试语料库，以支持许多机器学习和统计模式识别实验。

Proposal: 0326404PI: Lynette HirschmanTitle: Evaluation Bioinformatics TechnologyAbstractThe focus of this effort is to put in place an international "challenge evaluation" for text data mining, modeled after the successful CASP evaluations (Critical Assessment of Techniques for the Protein Structure Prediction). An evaluation will be implemented focused initially on two tasks: identification of biologically significant entities; and extraction of relation information such as interactions or identification of Gene Ontology terms in the literature. A gold standard of training and test corpora will be obtained by using expert annotated data from BIND, SWISS-PROT and other model organism databases. These data sets will be made available to a wide community through a website. Organizations in computational linguistics and information retrieval will also be involved. The overall strategy is to focus on problems of importance to working biologists, such as improved curation tools for biological databases, and better access to textual information in both the literature and in curated databases. This approach provides infrastructure and support to multiple research groups. It also supports creation of training and test corpora to support many machine learning and statistical pattern recognition experiments.