Consumer electronics, with 62 million tons of electronic waste (e-waste) generated in 2022 and e-waste expected to grow to 82 million tons annually by 2030, pose critical challenges when it comes to national infrastructure and circular economy policies. This paper compares forecasting approaches using sparse panel data for 32 European countries (2005–2018, Eurostat/Waste Electrical and Electronic Equipment (WEEE) Directive), focusing on leakage-safe prospective validation to guarantee true predictive performance. We make one-step-ahead predictions with conservative features (primarily lagged values) to account for temporal autocorrelation but with reduced multicollinearity (Variance Inflation Factor (VIF) ≈ 1.0). Cross-paradigm comparisons such as time-series baselines Autoregressive Integrated Moving Average (ARIMA), Seasonal ARIMA (SARIMA), Long Short-Term Memory (LSTM), hierarchical mixed-effects models, pooled machine learning (9 methods), and block-bootstrap-augmented stacking ensembles demonstrate stacking’s effectiveness, with a weighted validation R² of 0.992 for held-out 2017–2018 data. Time-series approaches demonstrate negligible predictive power (mean R² = −9683) given non-stationarity and limited samples, while the hierarchical approach provides virtually no benefit (Intraclass Correlation Coefficient (ICC) 0.011) amidst computational instability. Bootstrapping improves high-variance tonnage forecasts (Root Mean Squared Error (RMSE) reductions of 18.6%) while being detrimental to stable units, thus reinforcing parsimony. Feature ablation validates that only a few lags are necessary, preventing leakage from rolling means or calendar trends. Our method enables conservative year-ahead forecasts with quantified uncertainty, conservative estimates of e-waste management, allowing for buffer planning and policies even when data is scarce. By using strictly out-of-sample tests rather than biased ones, this work characterizes achievable year-ahead performance under sparse annual panels.

Accurate capacity and State of Charge (SOC) estimation are crucial for ensuring the safety and longevity of lithium-ion batteries in electric vehicles. This study examines ten machine learning architectures, Including Deep Belief Network (DBN), Bidirectional Recurrent Neural Network (BiDirRNN), Gated Recurrent Unit (GRU), and others using the NASA B0005 dataset of 591,458 instances. Results indicate that DBN excels in capacity estimation, achieving orders-of-magnitude lower error values and explaining over 99.97% of the predicted variable’s variance. When computational efficiency is paramount, the Deep Neural Network (DNN) offers a strong alternative, delivering near-competitive accuracy with significantly reduced prediction times. The GRU achieves the best overall performance for SOC estimation, attaining an $R^2$ of 0.9999, while the BiDirRNN provides a marginally lower error at a slightly higher computational speed. In contrast, Convolutional Neural Networks (CNN) and Radial Basis Function Networks (RBFN) exhibit relatively high error rates, making them less viable for real-world battery management. Analyses of error distributions reveal that the top-performing models cluster most predictions within tight bounds, limiting the risk of overcharging or deep discharging. These findings highlight the trade-off between accuracy and computational overhead, offering valuable guidance for battery management system (BMS) designers seeking optimal performance under constrained resources. Future work may further explore advanced data augmentation and domain adaptation techniques to enhance these models’ robustness in diverse operating conditions.