Why R? 2021 | Keynote | Feature-based Time Series Forecasting

Website - http://2021.whyr.pl/ Slack - http://whyr.pl/slack/
Speaker: Dr Thiyanga S. Talagala
Click Show More for bio + abstract

Abstract:
Many applications require a large number of time series to be forecast. Providing better forecasts for these time series is important in decision and policy making. However, large scale time series data present numerous challenges in modelling and implementation due to the high dimensionality. It is unlikely that a single method will consistently provide better forecasts across all time series. On the other hand, selecting individual forecast models when the number of series is very large can be extremely challenging. In this talk, I will present a general framework for forecast model selection using meta-learning. We call this framework FFORMS (Feature-based FORecast Model Selection). The underlying approach involves computing a vector of features from the time series which are then used to select the forecasting model. The model selection process is carried out using a classification algorithm – we use the time series features as inputs, and the best forecasting algorithm as the output. Furthermore, we explore what is happening under the hood of the FFORMS framework and thereby gain an understanding of what features lead to the different choices of forecast models and how different features influence the predicted outcome. The proposed algorithm is implemented in the R package seer, which is available on CRAN (https://CRAN.R-project.org/package=seer).

Bio
Dr Thiyanga S. Talagala is a Senior Lecturer in the Department of Statistics, University of Sri Jayewardenepura, Sri Lanka. Thiyanga earned her PhD in Statistics from Monash University, Australia in 2019. She is a co-founders and a co-organizer of R-Ladies Colombo, a local chapter of the R-Ladies Global Organization. R-Ladies Colombo is the first ever local chapter in Sri Lanka. Thiyanga is strongly committed to develop open source software tools to facilitate reproducible research and have written several R packages on CRAN for solving various types of problems in statistical modeling and research related problems. Thiyanga’s research focuses on developing new statistical machine learning tools to help both practitioners and theoreticians make more open, explainable and reproducible data-driven discoveries.