Commit ef42830b authored by John Doe's avatar John Doe
Browse files

Minor tweaks to Main work-flow defaults

parent 7645024b
......@@ -501,8 +501,8 @@ write_evals <- function( hist_bins = 25, write_errors=T, Surrogate_performance_t
# - together with specified preprocessing methods (see header for method enum descriptions)
# - and column rounding specification. the rounding spec is important because otherwise You
# may end up trying to predict numerical effects in the model output instead of meaningful output
Main <- function(surrogate_types = c(),
preprocessing_ind = c(17,22,23,24,25,26)
Main <- function(surrogate_types = c()
,preprocessing_ind = c(17)
# set this to actual data str to supply loaded tables
,input_data = "dumps/amd_course_IN.txt"
,output_data = "dumps/amd_course_OUT.txt"
......@@ -532,20 +532,20 @@ Main <- function(surrogate_types = c(),
# of ensemble and will not predicted in output, but will be given as inputs in validation.
# the reasoning is that these parameters come from the couple model (flow?) and are thus
# external to chemistry - are therefore forced on the surrogate completely from the outside
,external=c("Time","Pressure")
,allow_neg_cols=c("Pressure","Charge")
,external=c()
,allow_neg_cols=T
#,multiround = c(9,6)
,multiround = c()
#,exclude_output_columns = c("e")
,train_para=T, run_para=F, preproc_para=T, use_cores = 0
,train_para=F, run_para=F, preproc_para=T, use_cores = 0 # 0 = auto
,tuner = F
# you can apply a rounding specification, but it usually does more harm than good
# you can apply a rounding specification, including for each output column, like tis
#,r_digits = list( C=f6,Ca=6,Cl=6,H=12,Mg=5,e=9,Calcite=6,Dolomite=6
# or just set it to Inf for no rounding
,r_digits = Inf
,write_model_data=T
,write_full_residuals=F
,write_filepath_prefix="models/"
#,r_digits =
#,r_digits = list( C=f6,Ca=6,Cl=6,H=12,Mg=5,e=9,Calcite=6,Dolomite=6 )
) {
# run only the preprocessing methods that have not been blacklisted
......
This diff is collapsed.
output,method,preprocessing,speed,method_realname,RMSE,RSS,SAD,MAD,MASE,AME,RSSQ,model_id,MASE_speed_score,min_MASE
medv,brnn,range,0.00200000000040745,Bayesian Regularized Neural Networks,4.31012040679091,2823.72496399739,388.342279993718,2.55488342101131,0.403242956571774,23.2704632450485,53.1387331802085,356,0.0365327946913391,0.403242956571774
medv,brnn,range,0.00199999999995271,Bayesian Regularized Neural Networks,4.31012040679091,2823.72496399739,388.342279993718,2.55488342101131,0.403242956571774,23.2704632450485,53.1387331802085,356,0.0341317955315049,0.403242956571774
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment