本研究探讨了K-Means聚类算法,在不同距离度量基础上对配对交易中两种期货合约的历史价差序列进行分类的应用。本文比较了欧式距离、曼哈顿距离、切比可夫距离和余弦相似度在价差序列分类中的应用效果。研究结果表明,相较于传统的欧式距...本研究探讨了K-Means聚类算法,在不同距离度量基础上对配对交易中两种期货合约的历史价差序列进行分类的应用。本文比较了欧式距离、曼哈顿距离、切比可夫距离和余弦相似度在价差序列分类中的应用效果。研究结果表明,相较于传统的欧式距离,余弦相似度能够更好地对价差序列进行聚类,在效果评测指标上表现更加优异。This study explores the application of K-Means clustering algorithm to classify the historical spread sequences of two futures contracts in paired trading based on different distance measures. This article compares the application effects of Euclidean distance, Manhattan distance, Chebyshev distance, and cosine similarity in price difference sequence classification. The research results indicate that compared to traditional Euclidean distance, cosine similarity can better cluster price difference sequences and perform better in performance evaluation indicators.展开更多
文摘本研究探讨了K-Means聚类算法,在不同距离度量基础上对配对交易中两种期货合约的历史价差序列进行分类的应用。本文比较了欧式距离、曼哈顿距离、切比可夫距离和余弦相似度在价差序列分类中的应用效果。研究结果表明,相较于传统的欧式距离,余弦相似度能够更好地对价差序列进行聚类,在效果评测指标上表现更加优异。This study explores the application of K-Means clustering algorithm to classify the historical spread sequences of two futures contracts in paired trading based on different distance measures. This article compares the application effects of Euclidean distance, Manhattan distance, Chebyshev distance, and cosine similarity in price difference sequence classification. The research results indicate that compared to traditional Euclidean distance, cosine similarity can better cluster price difference sequences and perform better in performance evaluation indicators.