A stock data analysis method based on price linkage network

Abstract

The invention discloses a stock data analysis method based on a price linkage network. The method comprises the following steps that (1), stock price data is collected, the linkage relationship among the stock prices is calculated, and the price linkage network is built by using the stocks as nodes and the linkage relationship as the edge; (2), in the price linkage network, the upvaluation expectations of the stock nodes are calculated according to the parent node set in two hops of the stock nodes, the stock price trends and the linkage relationship weight; and (3), the stocks are sequenced according to the upvaluation expectations. According to the method, the linkage relationship among the stock prices is sufficiently excavated, the upvaluation expectations of each stock in the recent period can be reasonably judged according to the price fluctuation condition of the stock market, and the fair game problem possibly met in the single stock price trend prediction can be effectively avoided. The method has the characteristics that the calculation is simple, the timeliness, the flexibility and the expansion performance are realized, the processing quantity on the stock historical data is small, and the method is applicable to the stock market with the characteristics of great data quantity and frequent price fluctuation.

Description

technical field [0001] The invention relates to a data analysis method, in particular to a stock data analysis method based on a price linkage network, which is used for predicting the trend of stock prices. Background technique [0002] Investing in stocks requires making reasonable predictions about the direction and likelihood of future movements in stocks. In the stock market, there are various factors that can affect the stock price, including the economic environment, national policies, buyer psychology, etc., making it difficult to accurately grasp the characteristics of stock price trends. Stock trading has accumulated a large amount of historical data. The common method is to analyze and process the historical data, mine the correlation between the data, find out the law of change and establish a mathematical model, and on this basis, predict the trend of the stock price. [0003] Existing stock price data forecasting methods can be divided into two categories: one...

AI Technical Summary

Problems solved by technology

If there is a fair game, it is difficult to accurately estimate the future trend of a stock based only on the historical price series of a single stock

At present, stock data prediction methods based on association rules have been more and more researched and applied, but the existing methods are too computationally complex and lack the necessary screening of historical data; thus, it is difficult to respond to the ever-changing stock market in a timely manner and is not applicable Stock forecast requirements for short-term investment

Images