| DC 欄位 |
值 |
語言 |
| DC.contributor | 機械工程學系 | zh_TW |
| DC.creator | 林似霖 | zh_TW |
| DC.creator | Shih-Lin Lin | en_US |
| dc.date.accessioned | 2009-7-26T07:39:07Z | |
| dc.date.available | 2009-7-26T07:39:07Z | |
| dc.date.issued | 2009 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:444/thesis/view_etd.asp?URN=93343043 | |
| dc.contributor.department | 機械工程學系 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 數據分析是不可或缺，因為它是研究過程的一個重要階段，我們可以分析數據有新的發現。在這論文中有兩種好的數據分析方法，一種是獨立成份分析(ICA)，另一種是經驗模態分解(EMD)，本論文分成幾個部分，第一部分是應用獨立成份分析改善胎兒心電圖，經由獨立成份分析方法改善的結果可以更了解胎兒心跳的情況，第二部分是應用獨立成份分析於通訊保密上，由於獨立成份分析在估測的過程中可能會有失真與誤差的產生，所以估測出的結果會有相位相反和振幅大小不相等的現象，使得通訊保密還原的訊號失真，本研究中提出改良式的獨立成份分析，改良的獨立成份分析可以估測出真正的相位和振幅，使得保密的訊號可還原與原本一樣的訊號，第三部分是解決雜訊干擾問題，當我們要分析的數據受雜訊嚴重干擾，而且我們又無法用很多感知器去量測的數據分析問題，在此提出獨立成份分析和經驗模態分解結合的方法，當數據是高雜訊的情況時，利用兩個感知器可以分離出多個原始訊號，此方法是先用獨立成份分析進行處理，把雜訊和原始混合的訊號分開，再將原始混合的訊號進行經驗模態分解的分析，結果顯示此方法確實改善雜訊干擾問題。
| zh_TW |
| dc.description.abstract | Data are the only link we have with unexplained reality; therefore, data analysis is the only way through which we can find out the underlying processes of any given phenomenon. One of the most important goals of scientific research is to understand nature. Data analysis is a critical link in the scientific research cycle of observation, analysis, synthesizing, and theorizing. There are two commonly used data analysis methods. One is independent component analysis (ICA), and the other is empirical mode composition (EMD). However, it is difficult to conduct blind source separation (BSS) during data analysis. The first issue that needs to be dealt with is than the signals separated by traditional ICA shows opposite phase and unequal amplitude, leading to aliasing after the original signals are retrieved. The second major problem occurs when the number of sensors is greater than or equal to the number of sources; blind source separation becomes a difficult part of the underdetermined problem. These two types of problems have routinely been considered as an obstacle for source separation. The aim of this study is to solve two BBS problems. An algorithm method is proposed which can improve these problems. The modified ICA algorithm has applications in many different fields, such as for fetal electroencephalograms (EEGs) secure communications in chaotic systems, and chaos control in communications. Here the combined ICA-EMD is applied to low SNR simulated data, low SNR length-of-day data analysis and secure communications in chaotic systems. It is demonstrated that the combination of ICA and EMD can achieve better results. The two methods complement each other. The results show that these methods are an effective data analysis tool and have great potential for application in many different fields.
| en_US |
| DC.subject | 盲源分離 | zh_TW |
| DC.subject | 獨立成份分析 | zh_TW |
| DC.subject | 經驗模態分解 | zh_TW |
| DC.subject | 數據分析 | zh_TW |
| DC.subject | Blind Source Separation | en_US |
| DC.subject | ICA | en_US |
| DC.subject | EMD | en_US |
| DC.subject | Data analysis | en_US |
| DC.title | 數據分析中盲源分離之研究 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.title | Study on Blind Source Separation during Data Analysis | en_US |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |