| There are three kind of investors: hedgers, arbitragers and speculators in the capital markets. Most investors are speculators to earn spreads. A large number of speculators with different abilities to collect information hold different views about the securities, which caused the deviations between value and price of the securities. And statistical arbitrage is the main force of correcting securities mispriced. This paper devotes to providing reasonable and effective arbitrage strategy, so as to improve the efficiency of the allocation of financial resources in securities markets.In this paper, The Independent Component Analysis is used to reduce the dimension of the stock data. At the same time, We extract the characteristic values for clustering analysis. In the process of clustering, we present a method named silhouette value to judge the clustering results. We also choose the weighted stock with some fundamental data, for example, the sector and the market value.Then we use minimum tracking error model to get the stocks’ weight. All of the three sets of arbitrage gains are considerable and steady. And the maximum one can reach 66.34%. In this paper, we also compare the results of this statistical arbitrage strategy with the ETF one. The results show that the trading strategy in this paper is far better than the arbitrage trading strategy used by the general organizations.The innovation of this paper can be concluded into the following four points: using independent component analysis model before the clustering analysis, reducing the sensitivity about the abnormal points and dimensions of stock data, so that it improves the accuracy and efficiency of clustering analysis; using silhouette value as the criteria of clustering results; adding to the sectors and market value when chooses the weight stocks increases the diversity and persuasion; calculation of spread and arbitrage with the rolling window makes the arbitrage arbitrage process closer to the reality. |
PDF Full Text Request