From: Forecasting worldwide empty container availability with machine learning techniques
 | Mean absolute error | |
---|---|---|
Approach | Qingdao | Tokyo |
Neural Network | 0.0160 | 0.0499 |
Probabilistic Model | 0.0460 | 0.0712 |
Benchmark Naive | 0.0426 | 0.0680 |
Benchmark SES | 0.0202 | 0.0752 |
Total container | 12,729 | 1,843 |