Member-only story
6 Time Series Predictive Tasks You Should Know About
How different problems arise in time series and the challenges associated with them
In this post, I will describe 6 predictive tasks that can be put together with time series data sets. Each of these tasks have lots of relevant applications. So, if you are a data scientist working with time series data, these may be helpful to extract value from your data. By predictive I mean tasks where the goal is to predict a future state of the time series, or a current state which is not easily observable.
Here’s a brief outline for each task:
- Time Series Forecasting: Predicting the future values of a time series;
- Spatio-temporal Forecasting: Similar to time series forecasting but for several locations or trajectories;
- Exceedance Forecasting: Forecasting whether the upcoming values will exceed a pre-defined threshold;
- Anomaly Detection (a.k.a. Activity Monitoring): The timely detection of rare but disruptive events requiring action;
- Time Series Classification: Classifying time series into predefined classes;
- Survival Analysis: Predicting the time until an event of interest occurs.