Integrated methodologies (machine learning, data analysis, natural language) and their use in health/language applications and assistive technologies for people with disabilities

我的主要研究兴趣包括两个广泛的研究领域：(1)机器学习和 用于健康和语言应用的相关技术；以及(2)为有以下疾病的人提供的辅助技术 残疾人士。理论和实践的明智结合产生了理论和方法上的进步。 从而实现了高级原型软件实现。 这种理论/实践混合体的例子包括为循证医学寻找最佳证据， 用多段贝叶斯网络和混合方法验证独立积累的证据 系统设计采用机器学习等任务进行分类和预测。另一个示例使用Head 用于语音韵律、语言翻译、文本挖掘和n元语法的驱动阶段结构语法方法 分析以将语言研究概括到更广泛的问题类别，例如作者归属、分类 从患者-照顾者记录中检测轻度认知障碍患者的痴呆症水平 电子邮件中的恶意代码，以及设计帮助残疾学生的软件。 机器学习和相关技术在健康和语言应用中的应用：我们的研究有效地利用了加拿大健康组织的一个主要问题的解决方案：为基于证据的最佳实践建议寻找最佳证据。一个长期目标是开发一个基于证据的管理和最佳做法建议的决策支持系统，方法是：(1)从文件中建立自适应信息提取、文本和数据挖掘的原型，作为最佳做法的基础；(2)使用多段贝叶斯网络来推断概率解释，以验证建议的证据。HYRIS推广了我们的基于粗糙集的混合智能生存分析系统。 为残疾人提供的辅助技术帮助他们完成了其他方式无法完成或不易完成的任务。我们的研究专门为视力障碍、听力障碍、学习障碍和轻度认知障碍的人提供技术。

Two broad areas of investigation encompass my primary research interests: (1) machine learning and associated technologies for health and language applications; and (2) assistive technologies for people with disabilities. A judicious mixture of theory and practice has produced theoretical and methodological progress resulting in advanced prototype software implementations. Examples of this theory/practice admixture include finding best evidence for evidence-based medicine, verification of independently accumulated evidence using multiply sectioned Bayesian networks, and hybrid systems design using machine learning tasks such as classification and prediction. Another example uses head driven phase structured grammar approaches for speech prosody, language translation, text mining and n-gram analysis to generalize language research to a wider class of problems, e.g., author attribution, classifying dementia levels in patients with mild cognitive impairment from patient-caregiver transcripts, detecting malicious code in email, and designing software to assist students with disabilities. Machine learning and associated technologies for health & language applications: Our research effectively exploits solutions to a major problem for Canadian health organizations: finding best evidence for evidence based best practice recommendations. A long-term goal is to develop a decision support system for evidence based management and best practice recommendations by: (1) building prototypes for adaptive information extraction, text & data mining from documents on which to base best practices; & (2) employing multiply sectioned Bayesian networks to infer a probabilistic interpretation to validate evidence for recommendations. HYRIS generalizes our rough set-based hybrid intelligent system for survival analysis. Assistive technologies for people with disabilities help them accomplish tasks that they cannot accomplish otherwise or could not do easily. Our research specializes on technology for persons with vision impairment, with hearing impairment and with learning disabilities and mild cognitive impairment.