本研究以碎形特性為主題，針對受測者是否患有自律神經病變為實證對象，嘗試以不同的分析方法作為判斷的依據並對其結果進行評估。 所收集的實證對象分成健康人及重度自律神經病變糖尿病病人兩組樣本，再分別利用下列方式進行各種判斷指標的計算，這些判斷指標大致上可分為兩大類：一為以間隔時間為基準的指標，包括事件發生間隔時間折線圖及以發生間隔間隔時間為基準之週期圖；一為以發生事件次數為基準的指標，包括事件發生次數折線圖、variance-time curve、Fano factor及利用廣義發生率為基準之週期圖。由上述方式計算出判斷值之後，再利用統計檢定這些指標是否可區分出兩個樣本，進而評估哪些指標可用來作為判斷的參考依據。 經實證分析研究後所得到的結論為除了IBP與RBP之高低頻域比例不適用於判斷受測者是否患有自律神經病變外，其餘各種指標都可作為參考的依據，並達到可以提供不同於醫界所利用之方式來判斷受測者是否患有自律神經病變的目的。 According to fractal character, we take several analyses to calculate the indexes which attempt to check a particular patient suffers from autonomic neuropathy or not and also compare those indexes for different analyses. This research takes the pulse for serious diabetic autonomic neuropathy and healthy men as our empirical data. These analyses are based on interval and count measures. Interval measures consist of interveent-interval histogram and interval-based periodogram. And count measures consist of event-number histogram, variance-time-curve, Fano factor and generalized-rate-based peiodogram. Obeying the measures which is mentioned above, we can compute several indexes and use statistical test to verify these two different samples. Finally, we will review the conclusion and some suitable indexes will be chosen. The analysis we have conducted suggests the use of the conveniently indexes, unless the ratio of LF/HF for IBP and RBP, indicates whether a particular patient does or does not suffer from autonomic neuropathy.
