A non-conventional model to explain and forecast prices and trade flows of food staples across developing nations in Southeast Asia

本研究的目的是开发一个非传统的模型来理解和解释跨地理上不同的市场同质和易腐食品的贸易流。在处理来自发展中国家的大米价格数据时，我使用Web挖掘方法来获取尽可能多的数据源，并对数据进行数据清洗和可视化。然后，我采用了无监督的机器学习技术，如PrrinUNR成分分析（PCA）和t分布随机邻居嵌入（t-SNE），以更好地了解贸易伙伴群体之间食品价格的共同波动。这些数据压缩技术使我能够识别出食品价格高度联动的交易市场群体。

The purpose of this research is to develop an unconventional model to understand and explain the trade flows of homogenous and perishable food products across geographically distinct markets. When working with rice price data from developing countries, I used web mining methods to obtain as many sources of data as possible, and performed data cleaning and visualization of the data. I then employed unsupervised machine learning techniques such as Prrincipal Component Analysis (PCA) and t-Distributed Stochastic Neighbor Embedding (t-SNE), to better understand the co-movements of food prices across groups of trading partners. These data-compressing techniques allowed me to identify groups of trading markets whose food prices co-move to a high degree.