What is a prediction model?
A prediction model is a set of rules learned from past data. Learn what that means, what it needs, and why past patterns may not repeat.
When people say "AI predicts the price", they usually mean a model. A model is a set of rules, often a mathematical formula, that takes some inputs and produces an output. In forecasting, the inputs are things we know today (recent prices, how much the price has been swinging, trading volume) and the output is a guess about the future.
Most forecasting models are built by training: the computer looks at lots of past examples and adjusts its rules so its guesses would have matched what happened. If, in the past, a certain pattern was often followed by a price rise, the model learns to expect a rise when it sees that pattern again.
Pro unlocks all 72 Academy lessons, every quiz and the practice challenges. Six modules stay free forever.