marbin2050 / tfm_ntent Goto Github PK

DATA

It was employed the train data of FOLD1.

RESULTS

Summary of results [LINEAR REGRESSION]

Actual     Predicted

0 0 -4.884181e+07 1 1 8.145623e+06 2 0 1.079781e+07 3 1 -2.310273e+07 4 0 3.754552e+07 5 0 -4.289214e+06 6 1 6.988966e+06 7 0 5.392679e+06 8 0 -2.034758e+07 9 0 2.166298e+07 10 0 2.804398e+07 11 1 -2.237834e+07 12 2 -1.099401e+07 13 0 3.485450e+06 14 2 -8.666412e+06 15 2 8.909488e+05 16 2 -1.844772e+06 17 1 3.034339e+06 18 2 -2.115423e+07 19 0 2.923933e+06 20 0 -5.576761e+07 21 0 1.515343e+07 22 0 4.377544e+07 23 0 -2.190413e+07 24 2 2.896557e+07

Evaluation results [LINEAR REGRESSION]

Mean Absolute Error (MAE): 27042100.456406195 Mean Squared Error (MSE): 3.184783000488081e+16 Root Mean Squared Error (RMSE): 178459603.28567585 Spearman correlation: SpearmanrResult(correlation=-0.10286859411154027, pvalue=0.0) NDCG score: 0.8664693616694245

[LightGBM] [Info] Total Bins 2968 [LightGBM] [Info] Number of data: 413118, number of used features: 15 [LightGBM] [Info] Start training from score 0.678186

Summary of results [LIGHT GBM]

Actual  Predicted

0 0 0.656616 1 1 0.682010 2 0 0.702619 3 1 0.688516 4 0 0.635889 5 0 0.668084 6 1 0.690706 7 0 0.627553 8 0 0.686636 9 0 0.664848 10 0 0.673596 11 1 0.687345 12 2 0.685903 13 0 0.709910 14 2 0.675173 15 2 0.654059 16 2 0.712109 17 1 0.656542 18 2 0.682106 19 0 0.662230 20 0 0.694556 21 0 0.687678 22 0 0.665842 23 0 0.698781 24 2 0.627768

Evaluation results [LIGHT GBM]

Mean Absolute Error (MAE): 0.690847785394287 Mean Squared Error (MSE): 0.6711589292994014 Root Mean Squared Error (RMSE): 0.8192429000604163 Spearman correlation: SpearmanrResult(correlation=0.20098447722377355, pvalue=0.0) NDCG score: 0.9083823181462107

[LAMBDA RANK]

It is given an error with the whole data. It must be explored why.