Neural networks do not make any forecasts. Instead, they analyze price data and uncover opportunities. Using a neural network, you can make a trade decision based on thoroughly examined data, which is not necessarily the case when using traditional technical analysis methods.

This is an algorithm that uses a linear filter to add an object to a particular class or, on the contrary, to exclude it from the same object class. This is how the inequation looks like:

` w1 * a1 + w2 * a2 + ... wn * an > d,`

where:

wi – weighting coefficient with index i,

ai – numerical value of a sign with object’s index i,

d – threshold value that often equals 0.

If the left side of the inequation appears to be higher than the threshold value, then the object belongs to a specific class, if it is lower, the same does not apply. In case when the object classification implies a separation into two classes, a single-layer neural network is sufficient.

It may seem that the inequation used in a neural network is somehow similar to a “shamanic spell” in regards to weighting factors. In reality, this is not the case. The principle of neural network operation has a geometric meaning.

In fact, a plane is described geometrically as a linear equation. For example, in a three-dimensional space the plane equation concerning the coordinates X, Y and Z is the following:

A * X + B * Y + C * Z + D = 0

The coordinates of all points located on one side of the plane in this space satisfy the inequation:

A * X + B * Y + C * Z + D > 0

And coordinates of all points positioned on the other side of the plane satisfy the inequation:

A * X + B * Y + C * Z + D < 0

Thus, if a plane equation and any points coordinates are known to us, we can divide a set of all points in space into two sets of points separated by this plane.

Respectively, weighting coefficients in a neural network inequation are constants that define a certain plane equation in the multidimensional space of objects’ signs. By means of inequation we can accurately determine, whether these objects lie on one or the other side of the specified plane. For this purpose it is sufficient to locate the objects’ coordinates and, by substituting them in the equation of the plane, compare with zero.

**Which type of neural network is used by stock market indices?**

Kuo, Chen, and Hwang (2001) developed a decision support system through combining a genetic algorithm based **fuzzy neural network (GFNN) and ANN** for stock market.

**Can neural networks predict Forex?**

This paper reports empirical evidence that **a neural networks model is applicable to the statistically reliable prediction of foreign exchange rates**. Time series data and technical indicators such as moving average, are fed to neural nets to capture the underlying “rules” of the movement in currency exchange rates.

**Can RNN be used for stock prediction?**

According to the experimental results, **the CNN-LSTM can provide a reliable stock price forecasting with the highest prediction accuracy**. This forecasting method not only provides a new research idea for stock price forecasting but also provides practical experience for scholars to study financial time series data.

Stock price data have the characteristics of time series. At the same time, based on machine learning long short-term memory (LSTM) which has the advantages of analyzing relationships among time series data through its memory function, we propose a forecasting method of stock price based on CNN-LSTM. In the meanwhile, we use MLP, CNN, RNN, LSTM, CNN-RNN, and other forecasting models to predict the stock price one by one. Moreover, the forecasting results of these models are analyzed and compared. The data utilized in this research concern the daily stock prices from July 1, 1991, to August 31, 2020, including 7127 trading days. In terms of historical data, we choose eight features, including opening price, highest price, lowest price, closing price, volume, turnover, ups and downs, and change. Firstly, we adopt CNN to efficiently extract features from the data, which are the items of the previous 10 days. And then, we adopt LSTM to predict the stock price with the extracted feature data. According to the experimental results, the CNN-LSTM can provide a reliable stock price forecasting with the highest prediction accuracy. This forecasting method not only provides a new research idea for stock price forecasting but also provides practical experience for scholars to study financial time series data.