## Time Series Forecasting with STL

STL stands for “Seasonal and Trend decomposition using LOESS”. It is a versatile and robust method for decomposing time series. This method decomposes a time series into its three main components: Classical decomposition methods vs. STL Before carrying on with STL, let’s introduce traditional methods. Here you have an example Read more…

## Evaluation Metrics for Time Series Forecasting

Evaluation metrics, also known as performance measures or evaluative metrics, are quantitative measurements used to evaluate the performance and quality of a model or algorithm in solving a particular problem. It provides a standardized way to evaluate and compare different models and algorithms based on specific criteria. Error Metrics Error Read more…

## Step-by-Step Guide to Time Series Forecasting with SARIMA Models

We have talked about ARIMA and SARIMA models previously, however, we have never shown a real case step by step. Let’s first recap, to make sure we know what an ARIMA model is. ARIMA model ARIMA (Auto-Regressive Integrated Moving Average) is a popular time series forecasting model that combines autoregressive Read more…

## Theoretical Introduction to VAR Models for Time Series Forecasting

There are times when we need to forecast several variables at the same time. For these occasions, traditional methods such as ARIMA or Exponential Smoothing are not sufficient since they are univariate methods. Vector AutoRegression (VAR) is a statistical model for multivariate time series analysis and forecasting. It is used Read more…

## Time Series Forecasting with Facebook Prophet I

Part I: Trend modeling Classical time series forecasting techniques rely on statistical models that require a significant amount of effort to fine-tune and tailor to specific industry data. This often involves adjusting parameters to ensure accurate performance, which requires in-depth knowledge of the underlying models. Prophet is an open-source library developed Read more…

## Exponential Smoothing for Time Series Forecasting

Exponential Smoothing is a popular time series forecasting method used for univariate data. While other methods, such as ARIMA models, develop a model based on a weighted linear sum of recent past observations or lags, Exponential Smoothing models employ an exponentially decreasing weight for past observations. This weight is calculated Read more…

## Clean your Time Series data II: Remove seasonality and normalize data

In the first part of this series, we saw that cleaning the data is an essential step in the time series analysis process. This consists of the following substeps: The article will provide a general guideline for removing the seasonality and normalizing the time series data. Please remember that depending Read more…

## Clean your Time Series data I: Missing values and detrending

Time series data is a sequence of observations recorded at regular time intervals and is commonly used in various fields such as finance, economics, and engineering. However, time series data can often contain noise, outliers, missing values, and other anomalies that can affect the analysis and interpretation of the data. Read more…

## Exogenous variables in ARIMA models

ARIMA models are very powerful for forecasting time series data when this data is univariate. However, there is a type of ARIMA model that can also consider other variables. This type of model is called ARIMAX, which stands for “Auto-Regressive Integrated Moving Average with eXogenous variables”. ARIMAX is an extension Read more…

## Forecasting methods in Time Series

When working with time series data, it is important to assess the quality of a model in a way that accurately reflects real-world situations. Generally, the simplified process of building a Machine Learning model is the following: This is the way it is traditionally done in Data Science. However, when Read more…