Lilly Tech Systems
Learn Time Series Forecasting
Master the art and science of predicting the future — from classical ARIMA models and Facebook Prophet to deep learning with LSTMs and Transformers.
Your Learning Path
Follow these lessons in order, or jump to any topic that interests you.
1. Introduction
Time series concepts, stationarity, trend, seasonality, and the forecasting workflow.
2. Classical Methods
ARIMA, SARIMA, exponential smoothing, and statistical tests for time series.
3. Prophet
Facebook Prophet for business forecasting with holidays, changepoints, and multiplicative seasonality.
4. Deep Learning
LSTM, GRU, Temporal Convolutional Networks, and Transformers for time series.
5. Feature Engineering
Lag features, rolling statistics, Fourier terms, calendar features, and external regressors.
6. Evaluation
Time series cross-validation, MAPE, RMSE, SMAPE, and backtesting strategies.
7. Best Practices
Production deployment, monitoring forecast drift, ensemble methods, and real-world case studies.
What You'll Learn
By the end of this course, you'll be able to:
Classical Forecasting
Apply ARIMA, SARIMA, and exponential smoothing to univariate time series with confidence intervals.
Deep Learning Models
Build LSTM and Transformer architectures optimized for sequential prediction tasks.
Engineer Features
Create powerful lag features, rolling statistics, and calendar-based features for better forecasts.
Evaluate Properly
Use time-series-aware cross-validation and the right metrics to avoid misleading results.