Investment portfolio method using dimensionality reduction

Abstract

The invention discloses an investment portfolio method using dimension reduction. The method comprises the following steps: S1, obtaining stock information on a target security market; S2, obtaining aprice sequence of the stock; S3, using technical analysis to respectively calculate stock technical indexes of each stock; S4, performing dimensionality reduction on the stock technical index by using a dimensionality reduction technology; S5, calculating the similarity among the stocks according to the stock technical indexes of the stocks subjected to dimension reduction; S6, filtering weak correlation among the stocks by using the minimum spanning tree; S7, carrying out stock clustering by utilizing a Bond algorithm; and S8, selecting stocks with high earnings and small variance from different categories as seeds of the combination, and constructing a stock investment combination. By utilizing the method, the non-systematic risk of combination can be effectively reduced.

Description

technical field [0001] The invention relates to the field of financial technology analysis, in particular to a method of investment portfolio using dimension reduction. Background technique [0002] In the portfolio domain, dimensionality reduction is a particularly important step prior to the measurement of stock correlations. The foundational work in the field of investment portfolio is the theory put forward by Markowitz, which quantifies the relationship between income and risk for the first time. The core is to minimize risk at the same level of income or maximize income at the same level of risk. According to Markowitz Covitz's theoretical choice of stocks has a small correlation, and uncorrelated assets are conducive to risk diversification, but they are easily affected by many factors when calculating the correlation between assets, mainly including three aspects: [0003] 1) With the development of finance, more and more technical factors are applied in the investm...

AI Technical Summary

Problems solved by technology

[0003] 1) With the development of finance, more and more technical factors are applied in the investment field, resulting in higher and higher dimensions of analytical data. When estimating the covariance matrix between stocks, due to the existence of the high-dimensional curse phenomenon in reality, high Dimensional data is bound to affect the accuracy of estimation;

[0004] 2) Financial data is mixed with various noises, and the existence of noise will affect the accuracy of the data, thereby affecting the construction and optimization of investment portfolios;

Images