Machine learning is applied to share price analysis.
The share price histogram is divided into windows from which
the future predictions are made. The model uses tensorflow and keras.
The model accuracy can be over 90%
depending on how the initial conditions are set.

Enclosed we show an unique way to forecast the Bitcoin course 6h, 12h, 24, etc. hours ahead in real-time.
The algorithm is fully autonomous and it corrects itself when executing.
It uses public data from internet and the software
resources which are free of charge.
The algorithm learns in time and the results have better accuracy when time goes on.
The same algorithm structure can be used for the estimation of
other currencies and stock courses as well.

A more detailed description of the modeling and the results are found
here.

Machine learning in breaking of numbers

The product of the prime numbers is a Riemann zeta function
\begin{equation}
\zeta(2) = \prod_{p\hspace{1mm}prime} \frac{1}{1-p^{-2}} = \frac{\pi^2}{6}.
\end{equation}
Topologically \( \zeta(2) \) represents a surface area of a torus
\( \zeta(2) = (2\pi R)*(2\pi r)\), where \( R = 1/4 \) and
\( r = 1/6 \).

We apply machine learning and gauge theory to prime numbers.
Content coming soon here.

NLP analysis of the social media data

In this example we analyze the social media feeds and make
Natural Language Processing (NLP) analysis to find correlations in the Bitcoin share price.