|
|
本帖最后由 tinghuan 于 2010-9-15 10:51 编辑
大家好，很不容易通过这个论坛的身份验证呀！我本人在工业界的研究实验室里负责对外学术合作。
“eHealth”是我们最近在关注的一个方向。由于我所在的公司并不是以健康医疗相关的信息化为主业的，更多是平台级的开发。而其具体到“eHealth” 又是一个很大、很宽泛的话题。所以我们希望从自身的研究方向出发，积极寻找学术合作伙伴，共同做一些前沿性的探索。我列举一些我们现在想到的Topics, 期待论坛里的朋友多提宝贵意见。很希望能够了解 1） 与这几个Topics 相关的国内的研究状况如何？ 有哪些专家？2） 这些Topics 是否与国内的医学信息学研究方向有出入？如果有，那么现在国内最关注那些方向？与之相关的计算机技术有哪些？
Overview
In recent years, the quantity of digital health data has been growing exponentially, but at the same time, the availability and dissemination of health data for open research are limited. Analyzing and synthesizing this unprecedented volume of digital health data presents many challenges to computer science professionals, medical researchers and clinicians, but its successful handling can offer improvements in personal health management, clinical care, medical research and public health.
Advanced data mining technologies create opportunities to explore rich yet complex health data on a large scale. In order to understand the new challenges and opportunities of health data mining research, we hope to explore how data mining technologies can discover the hidden “gold” in large-scale health data with the ultimate goal of improving health.
We encourage collaboration embracing multiple groups across organizations or institutions, in particular, the collaboration and interdisciplinary graduate students training between computer scientists and health experts to interpret the health data and validate the development techniques.
Topic of Interest
• Creation of research health data warehouse
The lack of aggregated, clean and de-identified health data warehouse for the research community is the main constraint of health data mining. We are looking for projects taking initiatives to make research health data warehouse available with the ultimate goal of improving the community research quality. The topic of interests include but are not limited to: the reconciliation and integration of data from different sources; the cleaning and annotating of health data such as personal health monitoring data, electronic health records, lab results, research data, as well as other public sources and government data; and the aggregation of health data to foster related research.
• Unstructured health data information extraction
Most health data are in unstructured format. Medical literature and medical records are written in natural language. CTs/X-rays/MRIs are scanned as images.
Extracting features and structured information is the first step in the data mining process. A wide range of natural language processing, text mining, and image-recognition techniques can be applied for information extraction. We are seeking projects for information extraction from unstructured health data. Topics of interest include but are not limited to: applying NLP tools in multilingual medical-term recognition, ontology, and relation extraction with the objective to make the knowledge available and searchable; using text mining to understand the context, content, and scope of the text health data, in order to make connections and discover relationships; and utilizing computer vision and artificial intelligence techniques for image feature extraction.
• Machine learning on health data
Machine learning provides fundamental statistical-computational theories and algorithms of learning processes to discover hidden regularities in empirical data. A major focus of machine-learning research is to automatically learn to recognize complex patterns and make intelligent decisions based on data. A more recent trend is to observe that typical machine-learning approaches, such as classification, clustering, and regression, can be used to model and optimize the rich data generated in modern medical and health-care settings.
We seek projects to apply machine-learning methods and tools to medical and health-care applications. Topics of interest include but are not limited to: disease modeling and detection, patient monitoring and alarm systems, and prediction of treatment outcomes. |
|
发表于 2010-9-15 10:47:56
发表于 2010-9-15 23:22:37