Data Evolution and Data Variety in Cloud-Based Decision Making for Health Care Systems
讲座名称：
Data Evolution and Data Variety in Cloud-Based Decision Making for Health Care Systems
讲座时间：
2016-09-23
讲座人：
Hamido Fujita
形式：
校区：
兴庆校区
实践学分：
讲座内容：
应电信学院潘志斌教授邀请，Iwate Prefectural University, Japan的Hamido Fujita来我校交流并举行学术讲座，欢迎广大师生届时参加。
报告题目：Data Evolution and Data Variety in Cloud-Based Decision Making for Health Care Systems
报告时间：2016年9月23日(周五)，下午16:30-17:30
报告地点：西安交大兴庆校区西一楼538报告厅
报告人：Chair Professor Dr. Hamido Fujita教授（Iwate Prefectural University, Japan）
报告内容：In this keynote I will highlight the role of data variety and data evolution in decision making in the cloud, as big data in variety of forms have to be preprocessed balanced and customized to be used for accurate feature extract and data analysis in cloud to achieve good prediction with satisfactory accuracy. The medical decision support based on sampled data analysis in the cloud in relation to preferences collected from a situated environment could be inaccurate or unbalanced due big data evolution and uncertainty. In this keynote there is a need to highlight on such hot research topics that is part of big data analysis and sees what are the challenge and possible solutions.
I will approach in this talk on issues in ensemble learning and multi-classification techniques current state of art. Also we look to the objective criteria in decision making using these approaches. I will emphasize in my talk on subjective correlations among criteria providing better projection on objective data in relation to the situation that require accurate predictions, like in health care systems. Subjectivity would be examined based on correlations between different contextual learning structures that are reflecting the framework of personal context, for example in nearest neighbor based correlation analysis fashion or applied neural network and related ensemble learning techniques. Some of the attributes incompleteness also may lead to affect the approximation accuracy. Attributes with preference-ordered domain relations properties become one aspect in ordering properties in rough approximations. We will look to these new directions in health care domain, for early health risk predictions, and provide several demonstrations in Virtual Doctor System (VDS) developed by my group as a system assisting human doctor who is practicing medical diagnosis in real situation and environment. The interoperability is represented by utilizing the medical diagnosis cases of medical doctor, represented in machine executable fashion based on human patient interaction with virtual avatar as robot interaction resembling a real doctor based on machine learning developed by my group.
相关视频