DBLoss: Decomposition-based Loss Function for Time Series Forecasting

Jilin Hu (Inception Institute of Artificial Intelligence) · Bin Yang (Aalborg University) · Chenjuan Guo (Aalborg University) · Xiangfei Qiu (East China Normal University) · Xingjian Wu (East China Normal University) · Hanyin Cheng (East China Normal University) · Xvyuan Liu (East China Normal University)

dblossdecomposition-based lossdeep learning forecastingexponential moving averagesforward propagationinformed decision-makingloss function mean squared error model design performance improvement real-world datasetsseasonalitystate-of-the-art models time series forecastingtrend decomposition

Time series forecasting holds significant value in various domains such as economics, traffic, energy, and AIOps, as accurate predictions facilitate informed decision-making. However, the existing Mean Squared Error (MSE) loss function sometimes fails to accurately capture the seasonality or trend within the forecasting horizon, even when decomposition modules are used in the forward propagation to model the trend and seasonality separately. To address these challenges, we propose a simple yet effective Decomposition-Based Loss function called DBLoss. This method uses exponential moving averages to decompose the time series into seasonal and trend components within the forecasting horizon, and then calculates the loss for each of these components separately, followed by weighting them. As a general loss function, DBLoss can be combined with any deep learning forecasting model. Extensive experiments demonstrate that DBLoss significantly improves the performance of state-of-the-art models across diverse real-world datasets and provides a new perspective on the design of time series loss functions.