The Health Index Prediction Model and Application of PCP in CBM Wells Based on Deep Learning

Table 3

Evaluation index of the regression model.

Evaluation index	Calculation formula	Criteria
Mean absolute percentage error		The smaller the MAPE, the smaller the error
Mean absolute error		The smaller the MAD, the smaller the error
Root mean square error		The smaller the RMSE, the smaller the error, and the larger the RMSE, the larger the error
Theil’s inequality coefficient		The closer the TIC value is to zero, the higher the prediction accuracy will be; when it is equal to zero, it means 100% fitting
Decisive factor		is between 0 and 1; the larger the value, the better the model fitting