Today I saw a very interesting linear regression analysis method, which I learned about when I studied the Gaussian process regression when I was studying the normal distribution of numbers.
First of all, we know that market prices are not simply linear, and there is no function that can describe the price trend. Simply put, this is a probabilistic market (it is important to realize this!), and any market price trend is probabilistic. Technical analysis only provides a higher probability of market trends. It seems that predicting the market is almost completely impossible.
In fact, there are already some methods for estimating uncertainty, such as machine learning and statistical model building. Recently, some foreign traders have implanted statistical models into ChatGPT, which can also predict price trends very well. Among these prediction models, the most common one is the use of Gaussian process regression for estimation. Although it cannot be 100% accurate, it has a good reference value.
Gaussian process regression is widely used in uncertainty estimation of weather forecasting, disease prediction and diagnosis in the medical field, estimation of disease development trends, traffic flow control, etc.
Gaussian processes are essentially used to model the distribution of random functions and can be used to estimate the relationship between the target variable and the input features and provide uncertainty estimates.
In Gaussian process regression, it is assumed that the value of the target variable is generated by a random process, which is the Gaussian process. For each input feature, a probability distribution is obtained.
The specific calculation method is provided on Wikipedia.
Gaussian process regression predicts future market trends by learning the correlation between data, and it can also provide a good reference when put into the price market.
It can be seen that the Gaussian process can provide a clearer trend forecast, but using Gaussian regression forecast alone cannot tell the entry and exit points. The best method is to use it together with other indicators.
There is no problem in coordinating with EMA, MACD, etc.
In short, this can be regarded as a technical method that can lead the market and has good reference value.
