It looks interesting, what is the free AI predictive model based on, is the price data from 2013 or is there an adjustment to it.
Is there human control in re-correcting predictions to give better results, or is all data calculated as it is, regardless of its accuracy?
The speed of adjusting expectations caught my attention, which is good.
I bookmarked the site and will provide more details in upcoming replies.
Is there human control in re-correcting predictions to give better results, or is all data calculated as it is, regardless of its accuracy?
The speed of adjusting expectations caught my attention, which is good.
I bookmarked the site and will provide more details in upcoming replies.
Hi! Fair questions.
Firstly, its important to point out that we are not predicting price, we are predicting a range of volatility, essentially a probability band of where we think the max and min price for the token would be.
So, no, we are not claiming to be 98% accurate at predicting the precise price of Ethereum on a particular day. What we are good at is predicting when we think there is a higher chance that the price range might move.
If you hover over the historical chart title it should show you the formula used to plot accuracy. You might also find the tooltips on the chart to the left useful, it explains how our volatitlity range and relative value indicators are calculated (which is the data behind the historical chart we are discussing).
Here's a Medium article that i wrote about how the LSTM forecast model works. https://medium.com/gnylabs/introducing-fabian-gnys-new-machine-learning-engine-6f3fefd93839
