Table 1 Prediction performance of the machine learning-based methods in terms of mean absolute error for each of the motility values and the overall average.

From: Machine Learning-Based Analysis of Sperm Videos and Participant Data for Male Fertility Prediction

Classical Machine Learning Results
MethodProgressiveNon-progressiveImmotileAverage Mean Absolute Error
Baseline
ZeroR17.2607.86013.66012.927
Participant Data Only
Elastic Net15.1989.52513.44112.721
Gaussian Process15.5569.76213.47412.931
Simple Linear Regression15.4169.28113.60112.766
SMOreg15.3559.44112.95912.585
Random Forests13.3128.88611.90511.368
Random Tree17.80110.95214.98414.579
Tamura Image Features Only
Elastic Net14.4007.75012.19011.447
Gaussian Process13.2307.26011.92010.803
Simple Linear Regression13.5208.17012.69011.460
SMOreg13.2207.26011.92010.800
Random Forests13.5307.40012.06010.997
Random Tree18.7009.96016.52015.060
Tamura Image Features and Participant Data
Elastic Net14.1309.89011.75011.923
Gaussian Process13.70010.12011.46011.760
Simple Linear Regression13.94010.24011.41011.863
SMOreg13.71010.14011.46011.770
Random Forests13.51010.00011.34011.617
Random Tree18.66013.27016.96016.297
  1. The best performing algorithm in each category is in bold.