Can artificial intelligence predict the movement of stocks and financial markets

A recent study showed that artificial intelligence cannot be used in market investments
Stock markets are a very profitable and popular way to grow capital, and it is enough to know that more than $6 trillion is traded daily in Forex alone, according to what the Finance Magnates platform stated in a recent report.

If we add to this number what is traded in other central stock markets, such as the Wall Street, London, Tokyo and other markets, the number becomes truly enormous. In fact, the global stock markets are the factory of capital and the citadel of global capitalism.

With the tremendous technical development that our era is witnessing with the entry of the products of the Fourth Industrial Revolution - including algorithms, artificial intelligence and machine learning - into every field of life, the question that arises is: Can artificial intelligence be used to predict the movement of markets?

The answer is definitely yes. There has been increased reliance recently on the use of artificial intelligence to predict stock market movement, but couldn’t blind trust in the use of this intelligence lead to a loss of capital for traders and investors around the world? To what extent can we rely on this technology to predict stock movement with reliable accuracy?

Sohrab Mokhtari, Kang Qi Yin, and Jin Liu, who are professors of computer engineering at the University of Florida, discussed this matter in their research paper entitled “The Effectiveness of Artificial Intelligence in Predicting the Stock Market Based on Machine Learning,” according to what the University of Florida reported. Cornell University.

Stock market analysis
Before delving into the study, the researchers discussed the three main analysis methods used by investors around the world to anticipate market movement. The following two methods are used to analyze the movement of stocks in the stock market:

Fundamental Analysis: This financial school attempts to calculate the intrinsic value of a stock based on the company's revenues, profitability, degree of liquidity available to it, and operating efficiency. Ideally, if the intrinsic value is greater than the LTP, it should be bought, and if its intrinsic value is less than the LTP, it should be sold.

Technical Analysis: This methodology uses historical stock price data using a set of indicators - such as the Relative Strength Index (RSI), the Moving Average Convergence/Divergence (MACD), and the Money Flow Index ( MFI) - You try to know the market movement through it. If the technical analysis indicates that the stock price will rise, this will lead to making a buy order, and if the analysis indicates that the stock price will fall, this will lead to taking a sell order.

Stocks, the stock market and trading in them
The study discussed the main analysis methods used by investors around the world to anticipate market movement (Getty)
There are other models used in forecasting, including:

“Efficient Market Hypothesis” (EHM): This hypothesis suggests that the stock price moves in the direction of public sentiment which is a reaction to the latest economic and political news published in the world.

“Adaptive Market Hypothesis” (AHM): attempts to predict the direction of market movement using theories based on psychology.

Returning to the study that attempted to address the problem of the possibility of predicting stock market movement using artificial intelligence and algorithms based on machine learning, in order to do this, the researchers studied the two main schools used in analyzing market movement - namely the technical analysis method and the fundamental analysis method - and the effectiveness of the algorithms based on Machine learning in predicting the movement of financial markets using the two methods mentioned.

For this, previously labeled datasets were used to train prediction algorithms, and evaluation metrics were used to examine the accuracy of the algorithms in the prediction process. The results showed that the “linear regression” model followed by the researchers predicts the closing price significantly with a narrow margin of error when using the technical analysis method, while when using algorithms to predict the fundamental analysis method, the model was able to predict market movement with an accuracy of 76%.

Results
These results indicate that although AI can predict stock price trends or general sentiment about the movement of financial markets, its accuracy is not sufficient. Furthermore, while the Linear Regression model can predict the closing price with a reasonable margin of error, it cannot accurately predict the same value for the next business day, i.e. this forecasting ability is valid for only one business day.

Therefore, this AI-based model cannot be used for long-term investments. On the other hand, the accuracy of algorithms in predicting whether to buy, sell or hold a stock is not satisfactory enough, and can lead to a loss of capital.