Comparison of Univariate Time Series Prediction Methods

Date and Time: 
Tuesday April 9th 2019
CG Auditorium
Maggie Sleziak-Sallee
Time series forecasting is one of the most important data analysis techniques for extracting meaningful information to predict future behavior. The specific properties of a time series, such as its time component and underlying relationships, pose challenges in developing forecasting models. In addition to popular statistical methods in time series forecasting, such as regression or autoregressive integrated moving average (ARIMA), applications using machine learning algorithms such as Facebook’s prophet or recurrent Neural Networks, have provided new opportunities in the pursuit of more accurate forecast models. This study applies different types of prediction algorithms to a univariate time series. Both statistical and machine learning methods will be used and evaluated for their accuracy and forecast length. Challenges and opportunities will be discussed as well as potential next options to improve model results.
Speaker Description: 
Maggie has been working as a Software Engineer for the University Corporation for Atmospheric Research (UCAR) since 2002. As part of her duties, she oversees data transfer and processing operations for Radio Occultation satellite missions such as the COSMIC-1 mission launched in 2006. Maggie holds a Bachelor of Science in Computer Science, as well as a Master of Science in Data Science from Regis University. She also holds a Bachelor of Arts and a Master of Music degrees, and she enjoys playing the violin with her Gypsy Jazz quartet.
File Sleziak_SEA2019.pptx13.38 MB

Event Category: