應用希爾伯特-黃轉換法於菜價預測探討

目的：本研究取甘藍、蘿蔔兩項蔬菜價資料，應用希爾伯特-黃轉換（Hilbert-Huang Transform）法預測未來菜價趨勢。方法：以重要農產品葉菜類之甘藍，以及汛期期間做為替代作物之根莖類蘿蔔為主要研究對象，以高雄市公有零售市場價格資料，取楠梓市場以及高雄市公有零售市場價格均價之近3年甘藍與蘿蔔價格，運用希爾伯特-黃轉換（Hilbert-Huang Transform）預測菜價的未來趨勢建議。結果：希爾伯特-黃轉換法可提供菜價未來整體趨勢的預測參考。結論：本研究對於蔬菜價格的趨勢，以甘藍及蘿蔔為研究對象，取高雄市零售市場均價以及楠梓市場價格進行研究，採用希爾伯特-黃轉換（Hilbert-Huang Transform, HHT）進行分析，得到甘藍及蘿蔔價格區間預測。預測菜價區間與實際菜價結果比較顯示，蘿蔔價格的區間預測較甘藍價格區間預測為佳，顯見葉菜類的菜價較根菜類的菜價更容易受到非時序常態的因素影響，建議可在對應時序下提前相關政策介入以降低葉菜類菜價異常波動。

Purpose: The research discussed the problem of vegetable price fluctuation, use Hilbert-Huang Transform algorithm to predict the future price of cabbage and radish.Method: The research picks the latest 3 years cabbage and radish price from Kaohsiung government retail sale marketplace, 2015 and 2016 price as base; 2017 price as test data. Using Hilbert-Huang Transform algorithm to predict the future price of cabbage and radish are going up or going down.Result：It can provide some useful trends in future for vegetables price prediction using the Hilbert-Huang Transform algorithm.Conclusion：The study is vegetables price prediction, pick cabbage and radish as the object of the research and the price from Kaohsiung government retail sale marketplace to do the research, Using Hilbert-Huang Transform algorithm to do the analysis, The forcast price of the radish is better than cabbage, so obviously the price of the leafy vegetables are more easily affected by temporal factor, so suggest the related policy involve earlier to reduce leafy vegetables price abnormal fluctuation.