References for Time-Series Analysis and Time-Series Forecasting

a personal note on the challenging topics

Photo from Pixabay

Part 1: Traditional Approach and Facebook Prophet

• Sidebar — Criticism on Facebook Prophet
• Sidebar — Time Series Made Easy! Really?

Part 2: Machine Learning on Non-Window Data

• Case Study 2.1 — Predicting Energy Consumption based on XGBoost
• Case Study 2.2 — Predicting Bike Sharing Demand based on LightGBM

Part 3: Machine Learning and Deep Learning on Window Data

• Case Study 3.1 — LSTM and Time Series Forecasting
• Sidebar — Confidence/Uncertainty Intervals
• Sidebar — Do We Really Need Deep Learning Models?
• Case Study 3.2 — Transformer and Time Series Forecasting

Part 1: Traditional Approach and Facebook Prophet

A Time Series Lecture from CMU

How and Why Facebook Prophet

Is Facebook’s “Prophet” the Time-Series Messiah, or Just a Very Naughty Boy?

Time Series Made Easy! Really?

Part 2: Machine Learning on Non-Window Data

Case Study 2.1 — Predicting Energy Consumption based on XGBoost (timestamp features)

Case Study 2.2 — Predicting Bike Sharing Demand based on LightGBM (timestamp/holiday/workday/weather/lag features)

Time Series Forecasting with Supervised Machine Learning

Part 3: Machine Learning and Deep Learning on Window Data

Introduction to Forecasting from UBER

Case Study 3.1 — LSTM and Time Series Forecasting

How to Develop LSTM Models for Time Series Forecasting - Machine Learning Mastery

• Confidence Intervals for Machine Learning Classifiers
• Uncertainty Intervals in Facebook Prophet
• How to Generate Neural Network Confidence Intervals with Keras?

Do We Really Need Deep Learning Models for Time Series Forecasting?

Case Study 3.2 — Transformer and Time Series Forecasting