Glosarium AI
Kamus lengkap Kecerdasan Buatan
Simple Exponential Smoothing
Forecasting method that assigns exponentially decreasing weights to past observations, used for time series without trend or seasonality.
Holt's Method
Linear exponential smoothing technique that explicitly models the level and trend of a time series without seasonal component.
Holt-Winters Method
Advanced forecasting algorithm that simultaneously captures the level, trend, and seasonality of a time series through triple exponential smoothing.
Alpha Smoothing Constant
Parameter between 0 and 1 that controls the relative weight given to the most recent observation in calculating the series level.
Beta Smoothing Constant
Smoothing parameter that regulates the adjustment of the trend component based on recent changes observed in the time series.
Gamma Smoothing Constant
Exponential smoothing coefficient that determines the speed of adaptation of the seasonal component to changes in periodic patterns.
Level Component
Smoothed baseline value of the time series that represents the de-trended and deseasonalized average at each period.
Trend Component
Systematic growth or decline rate of the time series, captured by the difference between successive smoothed levels.
Seasonal Component
Recurring periodic pattern of the time series that repeats at fixed intervals, adjusted multiplicatively or additively according to the model.
Additive Method
Holt-Winters approach where the forecast is the sum of level, trend, and seasonality components, appropriate when variance is constant.
Multiplicative Method
Holt-Winters variant where seasonality is proportional to the level, used when the amplitude of seasonal variations grows with the trend.
Short-term Forecasting
Main application of exponential smoothing, optimized for near time horizons where recent patterns dominate predictions.
Brown's Method
Double exponential smoothing technique equivalent to Holt's method, using a single smoothing constant for level and trend components.
Damped Holt-Winters
Damped extension of the Holt-Winters method that gradually reduces the impact of the trend to avoid unrealistic long-term forecasts.
Seasonal Decomposition
Process of separating a time series into its fundamental components (trend, seasonality, residual) prior to smoothing.
Weighted Exponential Smoothing
Fundamental principle where weights decrease exponentially with the age of observations, ensuring recent data more significantly influence forecasts.