How to Win a Data Science Competition (Week3-1 part2)
Kaggle 機械学習
Published: 2019-06-23

Metrics optimization

学習目標

  • Describe the role of correct metric optimization method in a competition
  • Analyze new metrics
  • Create constant baselines
  • Recall the most important classification and regression metrics
  • Describe what libraries can be used to optimize a particular metric

Regression metrics review II

今回対象のメトリクス

  • ®MSPE, MAPE
  • ®MSLE

MSPE: Mean Square Percentage Error

MSPE=100%NN∑i=1(yi−^yiyi)2

最適な定数は加重平均

MAPE: Mean Absolute Percentage Error

MSPE=100%NN∑i=1|yi−^yiyi|

最適な定数は加重中央値

®MSLE: Root Mean Square Logarithmic Error

対数スケールで計算された RMSE

RMSLE=√1NN∑i=1(log(yi+1)−log(^yi+1))2

=RMSE(log(yi+1),log(^yi+1)=√MSE(log(yi+1),log(^yi+1)

対数空間で RMSE の定数を見つけ、対数空間から逆変換する必要がある

参考