Pair trading is a market neutral strategy commonly used in hedge funds. There are two main phases of the pair trading: pair formation phase and trading phase. Finding profitable stock pairs during the pair formation phase is an important issue. Previous one-to-one pair trading mostly limits the pair formation phase to a small number of stocks. In this paper, we make an innovative improvement by extending one-to-one pair trading to many-to-many pair trading. We propose a framework that involves association rule algorithms and the OPTICS clustering algorithm, and uses the bipartite graph partition algorithm to form many-to-many pairs. Experimental results show that the framework proposed in this paper is effective in selecting many-to-many pairs for pair trading and more trading opportunities are obtained compared with traditional pair trading.
|Title of host publication||Web and Big Data - 6th International Joint Conference, APWeb-WAIM 2022, Proceedings|
|Editors||Bohan LI, Chuanqi TAO, Lin YUE, Xuming HAN, Diego CALVANESE, Toshiyuki AMAGASA|
|Number of pages||9|
|Publication status||Published - 10 Feb 2023|
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
Bibliographical noteFunding Information:
This work was supported in part by the National Natural Science Foundation of China under Grant No. 61602149, and in part by the Fundamental Research Funds for the Central Universities, China under Grant No. B210202078.
© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
- Many-to-many pair trading
- Association rule mining
- Unsupervised learning
- Bipartite graph