應用支援向量迴歸方法建立台南地區登革熱疫情之壓力測試模型

Abstract

隨著科技的發展，人類的生活已大幅改善，但對於氣候變遷與其隨之而來的影響，我們仍無法徹底了解，環境也未能永續維持，因此生長環境逐漸變的嚴苛，在傳染病的部分也是如此，譬如登革熱的疫情至今仍無法獲得有效的控制。在台灣，近幾年來登革熱的疫情甚至有加劇的趨勢，不只是人數或者死亡率的提升，連帶還有疫情的擴散效應，造成登革熱災情的範圍逐漸往緯度較高的地區侵襲。因此，為了有效了解登革熱疫情，本研究的主要目的是針對台灣登革熱疫情較嚴重之台南市，利用主成分分析合併變數，建構一個登革熱疫情人數的預測模型，並在預測變數中納入考量人口流動因子，以準確預測未來可能發生疫情之風險，並提出模型，進行壓力測試，以找出疫情擴散之各變數的警戒水平。由於支援向量迴歸的方法是以支援向量機為理論基礎來建構預測模型，可以用於解決非線性問題，且對預測誤差有良好的修正能力，因此本研究使用支援向量迴歸方法來建構登革熱疫情之預測模型並進行壓力測試。本研究所提出之登革熱疫情之預測模型方法與壓力測試分析對登革熱疫情嚴重地區決策單位確有相當程度之幫助，其他登革熱疫區情況應可參考本研究建構其登革熱預測模型。

Human life has been improved rapidly with the development of science and technology, but we still cannot completely understand the impact of the climate change on human life. With the climate change, the living environment becomes worse and development of the infectious diseases spread faster than before. The dengue epidemic still cannot be effectively controlled. The fatality rate of dengue epidemic are increased rapidly in recent years. Moreover, the spread effect of dengue fever becomes significant in high latitude regions. Therefore, the main objective of this study is to utilize Support Vector Regression (SVR) to construct a prediction model for the dengue infections. In the prediction model, the migration factor is also considered to accurately predict the risk of the future outbreaks of dengue epidemic. The result of the stress test of the SVR prediction model is helpful for controlling the spread of dengue epidemic in the future. The result of this study can help government to make decision on controlling the epidemic of dengue.