ATLO-ML: Adaptive Time-Length Optimizer for Machine Learning -- Insights from Air Quality Forecasting

2025-10-07

AI Summary

In time-series forecasting, manually selecting input window length and sampling rate—often based on domain expertise—hinders adaptability to user-specified output horizons. Method: This paper proposes ATLO-ML, an Adaptive Time-length and Sampling-rate Optimization framework for multi-step forecasting. ATLO-ML is the first approach to enable end-to-end joint optimization of input parameters (window length and sampling interval) driven directly by the target output duration, via a closed-loop pipeline comprising preprocessing, modeling, and performance feedback. Contribution/Results: The framework demonstrates strong cross-dataset and cross-scenario generalization. Evaluated on the GAMS benchmark and real-world air quality data from a data center, ATLO-ML consistently outperforms fixed-input baselines, reducing average prediction error by 12.7%–19.3%. These results validate the effectiveness and practicality of output-horizon-driven time-series modeling.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Search and Optimization: Sampling/Simulation-based Search
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Accurate time-series predictions in machine learning are heavily influenced by the selection of appropriate input time length and sampling rate. This paper introduces ATLO-ML, an adaptive time-length optimization system that automatically determines the optimal input time length and sampling rate based on user-defined output time length. The system provides a flexible approach to time-series data pre-processing, dynamically adjusting these parameters to enhance predictive performance. ATLO-ML is validated using air quality datasets, including both GAMS-dataset and proprietary data collected from a data center, both in time series format. Results demonstrate that utilizing the optimized time length and sampling rate significantly improves the accuracy of machine learning models compared to fixed time lengths. ATLO-ML shows potential for generalization across various time-sensitive applications, offering a robust solution for optimizing temporal input parameters in machine learning workflows.

Problem

• Optimizes input time length and sampling rate for time-series predictions
• Enhances machine learning accuracy by adapting temporal parameters dynamically
• Validates adaptive optimization using air quality forecasting datasets

Innovation

• Adaptive system optimizes input time length and sampling rate
• Dynamically adjusts parameters based on user-defined output length
• Validated with air quality data to enhance prediction accuracy