model_performace.md

Model benchmarks

lstm-31-standard-scale.keras

100 samples

Mean mse: 1.3301364859153182
Mean mae: 0.7860621329590152

July

Mean mse: 0.80207726491494
Mean mae: 0.6404199945926666

500 samples

Mean mse: 1.8233533360285212
Mean mae: 0.9130220972483978

3 July

mean squared error: 0.2759982614713334
mean absolute error: 0.4586808333333332

4 July

mean squared error: 0.4029899189961678
mean absolute error: 0.4287443333333343

lstm-21.keras

100 samples (NOT the same samples)

Mean mse: 1.4389040921201766
Mean mae: 0.8418800602555275

July

Mean mse: 0.8785829137258339
Mean mae: 0.7045396378040314

500 samples (NOT the same samples)

Mean mse: 1.8686362317624847
Mean mae: 0.9302549582504667

3 July

mean squared error: 0.22099141302891678
mean absolute error: 0.3459289166666668

4 July

mean squared error: 0.5667596347340832
mean absolute error: 0.4497885833333329

lstm-32-standard-scale-16steps.keras

1000 samples

Mean mse: 0.18841170006495292 Mean mae: 0.3279478771710768

10000 samples

Mean mse: 0.26812188855268576 Mean mae: 0.33343772142978384

Early July

Mean mse: 0.6828215889899365 Mean mae: 0.413766446518898

lstm-33

Early July

Mean mse: 0.3580409782558534 Mean mae: 0.3967901562690735

lstm-34

1000 samples

Mean mse: 0.1590597045592595 Mean mae: 0.2507294583925512

Early July

Mean mse: 0.2884590114027321 Mean mae: 0.34162370598316194