Course Review: Udacity’s Time Series Forecasting

In order to improve my skills in data science, I am always on the lookout for free courses that I can take to help me to learn more about my profession. I have found Udacity’s free online course, Introduction to Time Series Forecasting, and have decided to write a post about what I have learned in the course:-

Lesson 1: Fundamentals of Time Series Forecasting

In this lesson, simple forecasting methods were discussed:-

1. Average method
2. Moving average method
3. Naive method
4. Seasonal naive method

More sophisticated forecasting methods include:-

1. Exponential smoothing

A time series has a trend cycle.

A time series can also encompass seasonality, which is repeating values over a period of time.

A cyclical pattern is when the time series experiences rises and falls that are not of a fixed period. Many people confuse cyclical patterns with seasonal patterns. Cyclical patterns are much harder to predict than seasonal patterns.

Lesson 2: ETS Models (Error Trend Seasonality)

To determine the ETS, use a time series decomposition plot.