Output from NEURAL VARMAX Search 

Case: test                                                                  
optimal objective value: 1.28256   
 MAX_s=8 MAX_r=8 YDIM=1 XDIM=0 TRUE_NOBS = 751 DATANOBS = 746 
CALCULATION_MODE=1 TRUE_FHORIZON=-5 FHORIZON_NNET = 5 VALID_RATIO=0.0000 SR_WEIGHT=0.5000 
ALPHA=0.8000 ETA=0.3000 USE_LASTw=0 NOISE_FLIP=1 FEEDBACK_FLIP=1 RANDOM_FLIP=0
RANDOMIZE_NET=1 BINARY_FLIP=0 ACTIVATION_MODE=5 SCALE=1
neural structure: ni=9 ni_noise=17 nf=4 ns=1 DATANOBS=746
CHANGE_MARGIN=0.0000001000 final CHANGE = 1.0000000000 MAX_k=10 LAST_ITERATION=10
AIC = 0.462956980597561 
S_R = 0.951647761316190 
HQ  = 0.651307025919848 

Error statistics for the observed period

Series:             1 
Mean          -0.0003 
Stdev          1.1333 
Skewness       0.2949 
Kurtosis       5.6372 
MAPE           3.8322 
RMSE           0.0415 
Runs          -0.2833 
ErrorAngle     0.9152 

Error statistics for the forecast period

Series:             1 
Mean           0.0385 
Stdev          1.4130 
Skewness      -1.0467 
Kurtosis       2.8382 
MAPE          23.0443 
RMSE           0.5655 
Runs           2.8402 
ErrorAngle     0.9407