由于当今世界的不断发展，经济水平不断的提升，人们可支配收入的不断增加，金融市场越来越多的投资者涌现。无论是金融机构还是个人投资者都需要对自己的投资组合进行配置。资产配置不合理，不仅仅会导致自身利益受损，金融机构出现资产危机，甚至会导致金融市场出现动荡。因此，在保证投资有效性的前提下，如何满足风险偏好，同时最大化收益，是个人投资者与金融机构关注的焦点。

随着金融市场的发展，大量金融产品的不断涌现，为了规避风险，金融机构资产种类的不断增加。然而，资产数量的增加导致风险与夏普值的估计出现较大误差，为了解决该问题，本文采用MAXSER (Maximum-Sharpe-Ratio Estimated and sparse Regression Estimate, 简写为MAXSER)方法进行研究。MAXSER资产组合是均值方差资产组合问题的延伸，该模型将优化问题等价转化成无约束条件的回归模型，在线性回归基础上加入了对系数的限制条件，使之成为稀疏回归，从而达到降维目的。

为了验证模型的有效性，本文采用了四种资产组合方式进行实证研究：等权重的资产组合，随机分配权重的资产组合，均值方差资产组合以及MAXSER资产组合。MAXSER相较于其他三种方法能够更好的解决多维的方差与均值的估计问题，能够满足均值方差的约束条件，同时也能够将风险控制在一定范围内。因此，MAXSER模型为多维估计问题的解决提供了一定的思路。

With the development of today's world, more and more investors in the financial market emerge. Both financial institutions and individual investors need to configure their investment portfolios. Unreasonable asset allocation will not only cause damage to investors’ own interests, but also cause asset crises in financial institutions and even lead to financial market turbulence. Therefore, on the premise of ensuring the effectiveness of investment, how to satisfy the risk constraint and maximize the return is the focus of individual investors and financial institutions.

Due to the development of the financial market, a large number of financial products continue to emerge, and most financial institutions not only hold single products, so the number of assets continues to increase, which leads to a large estimated error between the risk and the Sharpe Ratio. In order to solve this problem, this article uses the MAXSER method for analysis. The MAXSER portfolio is an extension of the Mean-Variance portfolio problem. This model converts the optimization problem into an unconstrained regression model and redefines the response variables. On the basis of linear regression, restrictions on the coefficients are added to make it a sparse regression, so as to achieve the purpose of dimensionality reduction.

In order to verify the effectiveness of the model, this paper uses four asset portfolios for empirical research: equal-weight asset portfolios, asset portfolios with randomly assigned weights, Mean-Variance asset portfolios, and MAXSER asset portfolios. Compared with the other three methods, MAXSER can better solve the problem of multidimensional variance and mean estimation, meet the constraints of mean variance, and also control the risk within a certain range. Therefore, the MAXSER model provides some ideas for solving multidimensional estimation problems.