Skip to main content

Table 2 Mean computation times (seconds) for forecasting models training and evaluation

From: Concept and benchmark results for Big Data energy forecasting based on Apache Spark

Tests Training Evaluation
MLR LASSO Random forest MLR LASSO Random forest
\(R_\text {1D}\) 0.014 (0.009) 0.407 (0.074) 1773.078 (18.274) 0.01 (0.008) 0.073 (0.026) 583.438 (16.824)
\(SC8_\text {1D}\) 0.208 (0.015) 0.223 (0.026) 22.122 (1.639) 0.192 (0.027) 0.188 (0.015) 20.905 (1.755)
\(SCl_\text {1D}\) 0.399 (0.022) 0.404 (0.040) 10.342 (0.202) 0.462 (0.029) 0.440 (0.020) 7.247 (0.199)
\(R_\text {1W}\) 0.07 (0.006) 2.7 (0.075) OoM 0.072 (0.007) 0.27 (0.021) OoM
\(SC8_\text {1W}\) 0.271 (0.013) 0.281 (0.013) 108.993 (6.285) 0.348 (0.022) 0.353 (0.012) 71.187 (7.316)
\(SCl_\text {1W}\) 0.390 (0.022) 0.409 (0.025) 28.418 (0.491) 0.475 (0.025) 0.476 (0.018) 14.416 (0.169)
\(R_\text {1M}\) 0.324 (0.056) 13.438 (0.247) OoM 0.267 (0.074) 1.346 (0.223) OoM
\(SC8_\text {1M}\) 0.627 (0.015) 0.638 (0.018) 454.759 (3.311) 1.058 (0.15) 1.074 (0.031) 243.008 (6.786)
\(SCl_\text {1M}\) 0.463 (0.021) 0.482 (0.032) 67.469 (2.125) 0.644 (0.042) 0.652 (0.036) 33.546 (1.908)
\(R_\text {6M}\) 1.862 (0.146) 115.960 (17.589) OoM 3.206 (0.311) 13.610 (11.116) OoM
\(SC8_\text {6M}\) 2.822 (0.060) 2.857 (0.077) 3101.202 (14.726) 5.420 (0.101) 5.457 (0.150) 1722.941 (36.603)
\(SCl_\text {6M}\) 1.133 (0.055) 1.115 (0.028) 357.388 (5.037) 1.747 (0.050) 1.757 (0.061) 161.561 (2.156)
\(R_\text {1Y}\) 4.061 (0.165) OoM OoM 6.903 (0.718) OoM OoM
\(SC8_\text {1Y}\) 5.588 (0.050) 5.604 (0.041) 6291.860 (43.074) 10.809 (0.049) 10.800 (0.053) 3246.918 (35.248)
\(SCl_\text {1Y}\) 1.934 (0.061) 1.903 (0.055) 784.018 (11.487) 1.934 (0.061) 1.903 (0.055) 304.891 (3.768)
\(R_\text {5Y}\) OoM OoM OoM OoM OoM OoM
\(SC8_\text {5Y}\) 41.464 (0.575) 42.520 (1.937) NT 72.940 (1.726) 73.728 (2.644) NT
\(SCl_\text {5Y}\) 16.104 (0.711) 15.474 (0.683) IOF 26.528 (1.346) 26.380 (1.000) IOF
\(R_\text {10Y}\) OoM OoM OoM OoM OoM OoM
\(SC8_\text {10Y}\) NT NT NT NT NT NT
\(SCl_\text {10Y}\) 38.997 (1.869) 39.300 (1.755) IOF 63.044 (2.157) 63.127 (2.301) IOF
  1. The values in parenthesis are the standard deviation values
  2. Italics: lowest computation time for a given data mining technique and a certain amount of training/evaluation data
  3. OoM out of memory, NT not tested, IOF integer overflow