時 間：97年11月24日(星期一) 10:00 – 12:00
演講人：四維講座-Norden E. Huang 黃鍔院士
（Research Center for Adaptive Data Analysis National Central University 中央大學數據分析方法研究中心）
題 目： Data and Data Analysis in the New IT World新資訊科技世界的資料和資料分析
地 點：商學院六樓元大講堂
綱要： With the dawn of the new IT world, we will find data everywhere. In fact, we are practically drowning in data. Data are the only connect we have with the reality; therefore, data analysis is the only way we can find out the truth. Consequently, with the increasing volume of data, ‘Data analysis’ should be very different from the traditional ‘Data Processing’, for ‘Analysis’ implies taking the data apart and examination the components critically; while as ‘Processing’ means only going through a given set of algorithms. Examples of analysis are given first by textural categorization, which literally took the plays by William Shakespeare and Christopher Marlowe apart to the individual level and study the interrelationship among the components. Such examinations helped to settle the authorship of famous plays by Shakespeare and Marlowe. Examples of this technique applied to bio-medical data in disease diagnosis and DNA differentiation will also be given.
Categorization is useful, but it does not provide a quantification of the complexity within the data. As biological systems have to adapt to their complex environments, this evolution process makes all biological system extremely complex. It should be noted that complexity is not identical to randomness, which can be measured by entropy. To provide a measure for complexity, the Multi-Scale Entropy (MSE) method is introduced, which gives the means of separating healthy from the unhealthy quantitatively. Additional examples will be given on the heart failure and the balance measures.
Powerful as the MSE method is, stationarity is an implied requirement for the method to work properly. To overcome this hurdle, the HHT (Empirical Mode Decomposition, EMD, and Hilbert Spectral Analysis) is introduced to remove any trend in the data. HHT is an adaptive and versatile method for data analysis. Examples of using HHT are given for studying climate trends and cycles. Other applications of adaptive data analysis covering speech analysis, engineering and sciences problems will be discussed briefly during the talk.
As the natural processes are mostly nonlinear and nonstationary, adaptive data analysis methods are the only logical way to analyze our data and to further our understanding the underlying physical processes. With the variety of data from different processes, being adaptive is critical. As human, we evolved and survive because we are adaptive. So, we should learn to be adaptive in our treatment of data in this new IT World.
演講者介紹:
赫伯特－黃變換法」（Hilbert-Huang Transformation）重大成就發明人˙美國第一位獲得美國服務貢獻講殊榮華裔科學家
黃鍔院士因「赫伯特－黃變換法」（Hilbert-Huang Transformation，簡稱HHT)，在美國廣泛受到應用，並受美國航空及太空總署（NASA）推薦，榮獲「美國服務貢獻獎」（Service to America Medals）殊榮，該獎分為九大領域，黃鍔獲頒的領域為「科學與環境類」(Science and Environment)，成為第一位華裔科學家獲獎人。黃鍔院士在美國NASA工作超過三十年，為NASA 海洋科學首席科學家。曾於2000年、2003年以及2004年分別當選美國國家工程研究院院士、NASA年度發明家以及中央研究院院士。
HHT目前以美國 NASA 應用最多，包括利用來證實重力波實驗，進行宇宙的黑洞研究；研發飛機「變形翅膀」，取代飛機固定不動的雙翼；太空梭隔熱材料「保溫磚」穩固鑑定；衛星色譜分析，以了解植被生長速度，進而從事「溫室效應」研究。
除此之外，在國內黃鍔院士與中央大學防災中心合作，從事橋梁的安全檢測工作；與中研院、台大醫院神經科合作，進行生醫工程「肌肉控制」研究；並與中央大學大大氣系合作，進行全球氣候變遷觀測。
備 註：請至校園聯合系統報名