The drive to predict the stock market probably started at the same time as the stock market itself. But as you already know, now we can realize this "dream" with the help of predictive models.

Datasets with stock market records imply an analysis of trends, cyclical fluctuations, and seasonality. For example, stock markets tend to perform well at the start of the year, as that is when many investors have fresh capital. Share prices may rise ahead of long weekends and three-day holidays. This is due solely to human factors.

Basically, to predict the stock of the market, models are used that work with data that has multiple seasonality. One of the most popular models is Prophet, which was created by Meta. The mathematical model looks like this:

The equation includes the parameters of trends g(t), seasonality s(t), holidays h(t) and noise e(t)

You can experiment with the model in Python:

from prophet import Prophet

model = Prophet()
model.fit(df)