SSMF: Shifting Seasonal Matrix Factorization
This addresses forecasting challenges in dynamic, seasonal data streams like taxi demand, offering an online, automated solution, though it is incremental as it builds on matrix factorization with regime-switching enhancements.
The paper tackles the problem of forecasting future events in data streams with seasonal patterns that change over time, proposing the Shifting Seasonal Matrix Factorization (SSMF) method, which adaptively learns and switches between multiple seasonal regimes and outperforms state-of-the-art baselines on three real-world datasets.
Given taxi-ride counts information between departure and destination locations, how can we forecast their future demands? In general, given a data stream of events with seasonal patterns that innovate over time, how can we effectively and efficiently forecast future events? In this paper, we propose Shifting Seasonal Matrix Factorization approach, namely SSMF, that can adaptively learn multiple seasonal patterns (called regimes), as well as switching between them. Our proposed method has the following properties: (a) it accurately forecasts future events by detecting regime shifts in seasonal patterns as the data stream evolves; (b) it works in an online setting, i.e., processes each observation in constant time and memory; (c) it effectively realizes regime shifts without human intervention by using a lossless data compression scheme. We demonstrate that our algorithm outperforms state-of-the-art baseline methods by accurately forecasting upcoming events on three real-world data streams.