農業生產具季節性也受天候因素影響,因此農產品價格波動也比工業產品大。蔬果產品是民生必需品,其價格波動不但影響生產者,也影響消費者,所以農產品價格分析是長久以來受到重視的議題。結球白菜因產量多,在批發巿場交易量高,其價格高低影響市場蔬菜平均價格,是所謂的大宗蔬菜之一。台灣蔬菜生產主要集中於中南部地區,其中西螺地區是蔬菜集散地,西螺市場是具指標性的產地市場,故本研究以西螺市場為對象,採用類神經網路模式,以西螺市場在2011至20219年間結球白菜的交易價格資料,對資料進行訓練並根據訓練後的模型對2020年做預測,結果顯示使用LSTM模型,當日價格與前一日價格超過5%以上波動狀況時,預測結果準確度會失準,價格波動在5%以內的情況下,預測準確度相對較高。
Agricultural production will be affected by seasonality and climate. As a result, the price of agricultural products tends to fluctuate considerably more than industrial products. Vegetables are necessities and its price fluctuation will influence not only producer but also consumer so that price analysis of agricultural products has been highly concerned for a long time. Chinese Cabbage is one of the vegetables which are widely grown in Taiwan and its price will greatly impacts the average market price of vegetables due to higher yields and large trading volumes. Taiwan’s Vegetables production is mainly concentrated in the central and southern regions and the Xiluo area is a vegetable distribution center. Therefore, this study takes Chinese Cabbage in the Xiluo Agricultural Marketing Cooperation which is an indicative trading market as the object of study. Adopt LSTM networks to analyze trading data of the Xiluo Agricultural Marketing Cooperation for the period from 2011 to 2019 and train data then forecast the market price of Chinese Cabbage in 2020. The result shows that the predicted values will be inaccurate while the volatility rate of market price is over 5% compares to the previous day but represents a better accuracy while the volatility rate of market price is less than 5%.