What forecasting method uses historical data to predict future values by recognizing trends over time?

Time series forecasting is a method that focuses on analyzing historical data points collected or recorded at specific time intervals. By examining this data, it identifies patterns, trends, and seasonal variations, allowing for accurate predictions of future values based on the observed historical trends.

This method leverages the sequential nature of the data, which is crucial in understanding how a variable behaves over time. For instance, if a company has monthly sales data over several years, time series forecasting can highlight upward or downward trends, seasonality, and cyclical patterns that can influence future sales.

In contrast, other methods listed do not focus primarily on historical trends over time. Data cleansing involves preparing data for analysis by correcting or removing inaccuracies. Regression analysis, while it can involve time components, more broadly focuses on the relationship between variables rather than specifically the temporal trends inherent in time series data. Lastly, anomaly detection aims to identify unusual patterns or outliers in data rather than forecast future values based on historical trends.