Time Series Forecasting
What is Time Series Forecasting?
Time series forecasting consists of predicting future observations through the use of a model based on historical values and a time dimension. This additional time dimension provides a sequential structure for the data and is a source of information at the same time. Moreover, time series are often real-time, a feature that adds an error-reduction dimension to the model.
In classical statistics, the operation of predicting values through a time series is called extrapolation. This term explains the idea behind it, as the values of the times series are “extended” or “extrapolated” to forecast unknown future events.
Why are Time Series Important?
Time series are applied widely. They are used to forecast events in science, applied sciences, engineering, economics, agriculture, finance, health, and more. They have proven to be one of the most useful tools for forecasting events that occur sequentially in time.
Examples of its applications abound. From stock prices to product sales to birth rates in hospitals to labor demand, all use similar forecasting techniques, where future events are predicted based on historical data.
According to some authors, there is, however, a constraint to time series forecasting, because this approach considers only historical data and thus neglects unexpected events.
Time Series + LogicPlum
Time series is an important tool in modeling. However, handling them is not always easy, as they can present outliers, heteroskedasticity, lags, and autocorrelation. These problems can make the interpretation and forecasting of events difficult and demand expert knowledge in statistics and mathematics from researchers.
However, this difficulty is solved for LogicPlum’s users, as this platform performs all time-series operations in an automated manner without requiring technical knowledge from them. Additionally, the platform provides an automatically generated report that explains the model insights in a comprehensive manner.
As a result, LogicPlum’s users can focus on the model results, knowing that they are employing a prediction tool based on the latest statistical and mathematical developments.
Guide to Further Reading
For those interested in using time series with Python:
Prabhakaran, S. Time Series Analysis in Python – A Comprehensive Guide with Examples. Available at https://www.machinelearningplus.com/time-series/time-series-analysis-python/
For those who prefer R:
Srivastava, T. (2015). A Complete Tutorial on Time Series Modeling in R. Available at