What is time series?

A time series can be thought of as a list of data points ordered over a given time period. Data are kept at equal intervals. For a data set to be a time series, it must be time dependent. Data received one after another can affect each other.

There are many time series examples available. Some of these are weather records, economic data, the number of spots on the sun. Time series are used in fields such as statistics, forecasting (weather, etc.), signal processing. An example of a time series is shown below. We can see the values of one Ionosphere coefficient of the dataset I used in my thesis in a certain time.

  • Trend: It means that the values increase or decrease depending on time. The trend can be linear or curvilinear. There may also be series that do not show trends over time. An example of a linear trend is shown in the figure below. The image has an increasing trend over time.

  • Seasonality: Changes in a known frequency in certain periods. For example, the fact that the values increase in some seasons and exhibit similar behaviors at the same time intervals every year can be called seasonality. In the example below, each different colored line represents three consecutive years. It can be an example of seasonality as the values at the same time of each year show a similar pattern.

  • Cyclical fluctuations: If a time series has fluctuations longer than one year, it can be said to have cyclical fluctuations. Their movements are generally not predictive. This cyclical movement is sometimes called the "Work Cycle".
  • Randomness (Irregularity): Non-periodic irregular movements. It can occur randomly. The occurrence of these fluctuations is unpredictable or unpredictable. Not all components have to be in a time series. There are time series that can contain one or all of them.
  • Stationarity: Stationary, which is an important concept in time series, means that the statistics of a series do not change depending on time. If the mean, variance and covariance in a time series change with time, we can say that this series is not stationary. For example, "white noise", which is a kind of time series, is an example of a stationary series.

  • The purposes of a time series analysis may vary according to the fields. To give a few examples; In areas such as statistics, finance, meteorology, the purpose of time series analysis may be to make predictions. Time series analysis can be done for signal detection in control engineering. In areas such as machine learning, time series can be used for classification, anomaly detection, clustering or prediction.

Stay Informed!

By signing up for our e-bulletin, you can be informed about all our innovations.

"We use cookies to personalize and improve your Sisasoft Website usage experience. By making your visit with default settings, you accept the use of cookies as specified in Sisasoft's

Privacy Policy
0312 227 06 34