Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | 林育仕 | en_US |
dc.contributor.author | Yu-Shi Lin | en_US |
dc.contributor.author | 洪慧念 | en_US |
dc.contributor.author | Dr. Hui-Nien Hung | en_US |
dc.date.accessioned | 2014-12-12T02:47:24Z | - |
dc.date.available | 2014-12-12T02:47:24Z | - |
dc.date.issued | 2004 | en_US |
dc.identifier.uri | http://140.113.39.130/cdrfb3/record/nctu/#GT009226519 | en_US |
dc.identifier.uri | http://hdl.handle.net/11536/76889 | - |
dc.description.abstract | 一般的統計分析,偏重在兩變數間的相關性,而無法表示其因果關係。例如,身高跟體重有相當顯著的正相關,是身高影響體重還是體重影響身高呢?此例中,改變一個人的體重並不會使其長高或變矮,故體重不是影響身高的原因。但在現實生活、產業製程、甚至是生物基因科技中,有許多變數間的因果關係並不是那麼顯淺易見,所關心的變數之間到底何為親輩?何為子輩?一直是科學家所想要探索的議題。而基因網路正是目前分析變數間因果關係最強而有力的工具。 簡單來說,基因網路主要由兩部分所構成:一、代表變數的節點(node):二、連接節點之間的線段(edge),以箭頭表示因果方向(Lauritzen, 1982;Wermuth and Lauritzen, 1983;Kiiveri et al., 1984)。 結合專業領域知識與資料,有助於基因網路的建構。基因網路最重要的特色之一,便是具有學習功能。可藉由資料的不斷更新,來學習機率模型的設定,使推估達到穩定,有效的來作機率推論的工作(Pearl, 1986)。推論是基因網路最主要的用途,近年來其應用已相當廣泛,如醫學診斷、資訊檢索、預測市場未來走向、天氣預報和人工智慧等等。而目前最熱們的基因工程,探討某基因的存在是否會活化或抑制另一基因,基因間的因果關係更是科學家所想要迫切了解的(Pe'er et al., 2001 ; Spirtes et al., 2000 )。 本論文將應用兩種統計方法(PC-Algorithm & 貝氏計分法),探討變數間之因果關係並建構基因網路之模型。 | zh_TW |
dc.description.abstract | If we have several random variables, the statistician usually focus on the joint distribution between variables. But when two variables are highly correlated does not mean that one causes the other. In statistical term, we say that correlation does not imply causation. Over the last decade, researchers have been developed many methods for inferring causality. In this paper, we describe a Gene Network as an efficient tool for causality discovery by fusion of data and prior knowledge. A Gene Network is a graphical model that encodes probabilistic relationships among variables of interest. Recently, the Gene Network has become a popular representation for encoding uncertain expert knowledge in expert system. They have been used to aid in the diagnosis of medical patients and malfunctioning systems, to filter documents, to facilitate planning in uncertain environments. But the construction of a Gene Network can be a very difficult task to perform. For this reason, we discuss two methods: (1)PC-Algorithm, (2)Bayesian Score, for constructing Gene Network. | en_US |
dc.language.iso | zh_TW | en_US |
dc.subject | 基因網路 | zh_TW |
dc.subject | PC演算法 | zh_TW |
dc.subject | 貝式計分法 | zh_TW |
dc.subject | Gene Network | en_US |
dc.subject | PC-Algorithm | en_US |
dc.subject | Bayesian Score | en_US |
dc.title | 利用基因網路分析變數間之因果關係 | zh_TW |
dc.title | Causal Analysis with Gene Network | en_US |
dc.type | Thesis | en_US |
dc.contributor.department | 統計學研究所 | zh_TW |
Appears in Collections: | Thesis |
Files in This Item:
If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.