Predicting future values of sequential data (like currency prices) based on past observations. In forex, time-series forecasting uses statistical or machine-learning models (ARIMA, LSTM networks, etc.) to estimate future exchange rates or volatility. Forecasting models exploit temporal patterns, seasonality, and momentum in historical price series to make predictions. Accurate forex forecasting is crucial for robot strategies, as it guides decisions on entry/exit timing. Many AI-based forex robots apply time-series models (especially recurrent neural networks like LSTMs) to forecast short-term price movements and thereby automate trading decisions.