使用XGBoost识别时间序列中的结构

Using XGBoost to Identify Structures in Time Series

现实中的时间序列数据中一般包含杂讯而且维度很高。GBDT决策树算法在处理这类数据中有天生的优势。使用XGBoost来对股票数据的结构进行分类,结果显示了这种方法的有效性。

Real world time series data is notorious for its noise and high dimensionality. GBDT decision tree algorithm is naturally good at handling such data. we experimented this idea by using XGBoost on stock data,and results showed the effectiveness of this approach.