Gepsoft - GeneXproServer Knowledge Base

Home

About Us

Contact

Blog


What's New	Products	Buy Now	Downloads	Forum

GeneXproServer Online Guide Learn how to seamlessly integrate your models with your systems & workflow with the Online Guide.

Intellisense in Visual Studio

Editions

Last update: February 19, 2014

Settings

General Settings: Population Size and Model Structure; Sampling; Rounding Threshold; Simple Model; Structure Pressure; Fitness Function; Custom Fitness; Complexity Increase; Time Series Transformation; Random Numerical Constants; Genetic Operators; Genetic Operators - Random Constants; Other Settings
Linking Functions: Mathematical Functions; Boolean Functions
Simple Models (Fitness Functions)
Rounding Threshold Types
Fitness Functions: Regression & Time Series Prediction; Classification; Logistic Regression; Logic Synthesis
Random Numerical Constants Type
Favorite Statistics: Regression & Time Series Prediction; Classification & Logistic Regression; Logic Synthesis

General Settings

This is the list of settings that can be changed using the job files (settings node) in GeneXproServer 5.0. The values in the key columns are the identifiers that must be used in the "key" attribute of the setting node.

Note that not all settings' values are checked by GeneXproServer. It is advisable to test the values in GeneXproTools before using them in GeneXproServer.

Key	Data Type	Definition
Population Size and Model Structure
Chromosomes	int positive [1 4096]	Number of Chromosomes
HeadSize	int positive [1 256]	Head Size
Genes	int positive [1 64]	Number of Genes
LinkingFunction	List	Linking Function
Sampling
TrainingRecords	int positive	Size of the Training Set
TrainingSamplingType	int positive [0 10]	Sampling Type for the Training Set
ValidationRecords	int positive	Size of the Validation Set
TestingSamplingType	int positive [0 10]	Sampling Type for the Validation Set
MiniBatchMode	[0 1]	Turn on Mini Batch
MiniBatchModePeriod	int	Mini Batch Period value
Rounding Threshold
RoundingThresholdType	List	Rounding Threshold Type
RoundingThreshold	float	Rounding Threshold Value
LogisticRoundingThresholdBins	int	Number of Bins (Logistic Rounding Threshold)
Simple Model (Fitness Function)
SimpleModel	List	Fitness Function Reference Simple Model
PredictorVariable	variable index	Predictor Variable Index
Structure Pressure (Fitness Function)
SetParsimonyPressure	[0 1]	Turn on Parsimony Pressure
ParsimonyPressureRate	positive float [0 1]	Parsimony Pressure Rate
SetVariablePressure	[0 1]	Turn on Variable Pressure
VariablePressureRate	positive float [0 1]	Variable Pressure Rate
Fitness Function
FitnessFunction	List	Fitness Function
CostGainFN	float negative	False Negative Cost
CostGainFP	float negative	False Positive Cost
CostGainTN	float	True Negative Cost
CostGainTP	float	True Positive Cost
Increment	float	Increment Value
InnerMargin	float	Inner Margin Value
OuterMargin	float	Outer Margin Value
Reward	float positive	Reward Value
Punishment	float	Punishment Value
SelectionRange	float	Selection Range Value
Precision	float	Precision Value
Beta	float positive	Beta Value (F Measure Fitness)
LowerBound	float	Lower Bound
UpperBound	float	Upper Bound
Bins	int	Number of Bins (Fitness Functions)
Custom Fitness
CustomFitnessFunction	string	Custom Fitness Function Code
CustomFitnessTimeout	seconds	Custom Fitness Timeout Value
Complexity Increase
SetComplexityIncrease	[0 1]	Turn on Complexity Increase
ComplexityIncreaseNumberOfTries	int positive [1 32767]	Number of Tries
ComplexityIncreaseGenerationsWithoutChange	int positive [1 32767]	Number of Generations without Change
ComplexityIncreaseMaxComplexity	genes [1 64]	Maximum Complexity Value
Time Series Transformation
TimeSeriesDelayTime	int positive	Delay Time Value
TimeSeriesEmbeddingDimension	int positive	Embedding Dimension Value
TimeSeriesTestingPredictions	int positive	Number of Testing Predictions
Change Prediction Mode	--	--
Random Numerical Constants
UseRNC	[0 1]	Turn on Random Constants
ConstantsPerGene	int positive [1 100]	Number of Random Constants per Gene
DataTypeRNC	List	Random Constants Data Type (float/integer)
UpperBoundRNC	float	Random Constants Upper Bound
LowerBoundRNC	float	Random Constants Lower Bound
Genetic Operators
BestCloningRate	positive float [0 1]	Best Cloning Rate
BiasedLeafMutationRate	positive float [0 1]	Biased Leaf Mutation Rate
BiasedMutationRate	positive float [0 1]	Biased Mutation Rate
ConservativeFunctionMutationRate	positive float [0 1]	Conservative Function Mutation Rate
ConservativeMutationRate	positive float [0 1]	Conservative Mutation Rate
ConservativePermutationRate	positive float [0 1]	Conservative Permutation Rate
FunctionInsertionRate	positive float [0 1]	Function Insertion Rate
GeneRecRate	positive float [0 1]	Gene Recombination Rate
GeneTranspRate	positive float [0 1]	Gene Transposition Rate
ISRate	positive float [0 1]	IS Transposition Rate
InversionRate	positive float [0 1]	Inversion Rate
LeafMutationRate	positive float [0 1]	Leaf Mutation Rate
MutationRate	positive float [0 1]	Mutation Rate
OnePointRecRate	positive float [0 1]	One Point Recombination Rate
PermutationRate	positive float [0 1]	Permutation Rate
RISRate	positive float [0 1]	RIS Transposition Rate
RandomChromosomesRate	positive float [0 1]	Random Chromosomes Rate
RandomCloningRate	positive float [0 1]	Random Cloning Rate
StumblingMutationRate	positive float [0 1]	Stumbling Mutation Rate
TailInversionRate	positive float [0 1]	Tail Inversion Rate
TailMutationRate	positive float [0 1]	Tail Mutation Rate
TwoPointRecRate	positive float [0 1]	Two Point Recombination Rate
UniformGeneRecombinationRate	positive float [0 1]	Uniform Gene Recombination Rate
UniformRecombinationRate	positive float [0 1]	Uniform Recombination Rate
Genetic Operators - Random Constants
ConstantFineTuningRateRNC	positive float [0 1]	Constant Fine Tuning Rate
ConstantInsertionRateRNC	positive float [0 1]	Constant Insertion Rate
ConstantRangeFindingRateRNC	positive float [0 1]	Constant Range Finding Rate
PermutationDcRateRNC	positive float [0 1]	Dc Permutation Rate
DcInversionRateRNC	positive float [0 1]	Dc Inversion Rate
DcMutationRateRNC	positive float [0 1]	Dc Mutation Rate
DcTranspRateRNC	positive float [0 1]	Dc Transposition Rate
MutationRateRNC	positive float [0 1]	RNC Mutation Rate
Other Settings
Author	string	Author's Name
Comments	string	Run Comments
UseUDFs	[0 1]	Turn on UDFs
FavoriteStatistic	List	Set Favorite Statistic
Math & Boolean Functions	--	--

Linking Functions

The table below lists the values that can be used for the setting LinkingFunction. For example, the code below sets the Linking Function to Subtraction:

<setting key='LinkingFunction' value='1'/>

Linking Function	Value
Mathematical Linking Functions
Addition	0
Subtraction	1
Multiplication	2
Division	3
Avg2	98
Min	92
Max	95
Boolean Linking Functions
AND	10001
OR	10002
NAND	10003
NOR	10004
OR	10005
NXOR	10006
Less Than	10022
Greater Than	10023
Less Or Equal	10024
Greater Or Equal	10025

Simple Models (Fitness Functions)

The table below lists the values that can be used for the setting SimpleModel. For example, the code below sets the SimpleModel to Average Simple Model:

<setting key='SimpleModel' value='1'/>

And the code below shows how to set the PredictorVariable to d1 for the SimpleModel Predictor Variable:

<setting key='SimpleModel' value='3'/>
<setting key='PredictorVariable' value='1'/>

Simple Model	Value
Classification, Logistic Regression, Regression & Time Series Prediction
Target Average	0
Target Midrange	4
Target Median	5
Average Simple Model	1
Random Target Output	2
Predictor Variable	3
Logic Synthesis
Target Majority	6
Majority Simple Model	7
Random Target Output	2
Predictor Variable	3

Rounding Threshold Types

The table below lists the values that can be used for the setting RoundingThresholdType. For example, the code below sets the Rounding Threshold Type to Average:

<setting key='RoundingThresholdType' value='2'/>

Rounding Threshold Type	Value
User Defined	0
ROC	1
Average	2
Midrange	3
Midhinge	4
Class Proportions	5
Active Model	6
RMS	7
Interdecile Mean	8
Interquartile Mean	9
Trimean	10
Logistic	11
Large Margin	13

Fitness Functions

The table below lists the values that can be used for the setting FitnessFunction. For example, the code below sets the Fitness Function to Absolute Hits with RRSE:

<setting key='FitnessFunction' value='120'/>

Fitness Function	Value
Regression & Time Series Prediction
Relative Hits with RMSE	116
Relative Hits with RRSE	117
Relative Hits with RRSE 2	118
Absolute Hits with RMSE	119
Absolute Hits with RRSE	120
Absolute Hits with RRSE 2	121
Relative Error with SR	0
Relative Error/Hits	1
Absolute Error with SR	2
Absolute Error/Hits	3
Trends with Punishment	79
Bounded Trends	86
Trends R2 with Punishment	103
Bounded Trends R2	104
MSE	5
RMSE	6
MAE	7
RSE	8
RRSE	9
RAE	10
RSE 2	95
RRSE 2	96
RAE 2	97
Weighted MSE	80
Weighted RMSE	81
Weighted MAE	82
Weighted RSE	83
Weighted RRSE	84
Weighted RAE	85
Weighted RSE 2	98
Weighted RRSE 2	99
Weighted RAE 2	100
Enhanced MSE	28
Enhanced MAE	30
Enhanced RSE	31
Enhanced RAE	33
Enhanced RSE 2	101
Enhanced RAE 2	102
Quantile MSE	87
Quantile RMSE	88
Quantile MAE	89
Quantile RSE	90
Quantile RRSE	91
Quantile RAE	92
R-square	4
Bounded R-square	29
Positive Correl	26
Bounded Positive Correl	32
Unsupervised Free-form Features	128
Custom	13
Classification
ROC Measure	71
Bounded ROC	72
ROC with MSE	73
Rank Measure	74
Bounded Rank	75
Rank with MSE	76
Maximum Likelihood	112
Hinge Loss	124
Logistic RMSE	113
Dual Margin	24
Margin with Penalty	25
Triple Margin	49
Bounded Triple Margin	50
Enhanced Triple Margin	51
Bounded Enhanced Triple Margin	52
Entropy	17
Entropy with MSE	65
Purity	20
Purity with MSE	66
Positive Correl	15
Positive Correl with MSE	77
Bounded Positive Correl	27
Class Variation	126
MSE	53
RMSE	54
MAE	55
RSE	56
RRSE	57
RAE	58
RSE 2	105
RRSE 2	106
RAE 2	107
Enhanced MSE	16
Enhanced MAE	18
Enhanced RSE	19
Enhanced RAE	21
Enhanced RSE 2	108
Enhanced RAE 2	109
Hits with Punishment	11
Cost/Gain Matrix	23
Sensitivity/Specificity	12
PPV/NPV	14
SSPN	22
F Measure	122
Jaccard Similarity	93
Matthews Correl	67
Matthews Correl with MSE	68
False Balance	129
Enhanced False Balance	130
Relative False Balance	131
Enhanced Relative False Balance	132
Unsupervised Free-form Features	127
Custom	13
Logistic Regression
ROC Measure	71
Bounded ROC	72
ROC with MSE	73
Symmetric ROC	69
Bounded Symmetric ROC	70
Rank Measure	74
Bounded Rank	75
Rank with MSE	76
Quantile Spread	61
Bounded Quantile Spread	62
Quantile Spread with MSE	63
Maximum Likelihood	114
Hinge Loss	125
Logistic RMSE	115
Dual Margin	24
Margin with Penalty	25
Triple Margin	49
Bounded Triple Margin	50
Enhanced Triple Margin	51
Bounded Enhanced Triple Margin	52
Entropy	17
Entropy with MSE	65
Purity	20
Purity with MSE	66
R-square	59
Bounded R-square	60
Positive Correl	15
Positive Correl with MSE	77
Bounded Positive Correl	27
Class Variation	126
MSE	53
RMSE	54
MAE	55
RSE	56
RRSE	57
RAE	58
RSE 2	105
RRSE 2	106
RAE 2	107
Enhanced MSE	16
Enhanced MAE	18
Enhanced RSE	19
Enhanced RAE	21
Enhanced RSE 2	108
Enhanced RAE 2	109
Hits with Punishment	11
Cost/Gain Matrix	23
Sensitivity/Specificity	12
PPV/NPV	14
SSPN	22
F Measure	122
Jaccard Similarity	93
Matthews Correl	67
Matthews Correl with MSE	68
False Balance	129
Enhanced False Balance	130
Relative False Balance	131
Enhanced Relative False Balance	132
Unsupervised Free-form Features	127
Custom	13
Logic Synthesis
Hits with Punishment	34
Cost/Gain Matrix	35
Sensitivity/Specificity	40
PPV/NPV	41
SSPN	42
F Measure	123
Jaccard Similarity	94
Matthews Correl	44
Matthews Correl with MSE	64
Entropy	36
Entropy with MSE	38
Purity	37
Purity with MSE	39
Positive Correl	43
Positive Correl with MSE	78
Enhanced MSE	45
Enhanced MAE	46
Enhanced RSE	47
Enhanced RAE	48
Enhanced RSE 2	110
Enhanced RAE 2	111
Custom	13

Random Numerical Constants Type

The table below lists the values that can be used for the setting DataTypeRNC. For example, the code below sets the Random Numerical Constants Type to Floating-Point values:

<setting key='DataTypeRNC' value='1'/>

Data Type	Value
Integer	0
Floating-Point	1

Favorite Statistics

The table below lists the values that can be used for the setting FavoriteStatistic. For example, the code below sets the Favorite Statistic to RMSE:

<setting key='FavoriteStatistic' value='17'/>

Statistic	Value
Regression & Time Series Prediction
MSE	16
RMSE	17
MAE	18
RSE	20
RRSE	21
RAE	19
Correlation Coefficient	22
R-square	2
Outliers	15
Hits	37
Up/Down Accuracy	40
Up/Down Error	41
Up/Down Hits	42
Up/Down Errors	43
Classification & Logistic Regression
Accuracy	5
Error	6
Hits	38
Errors	39
Sensitivity	7
Specificity	8
Positive Predictive Value	9
Negative Predictive Value	10
Recall	33
Precision	34
F1 Measure	35
Jaccard Similarity	31
Matthews Correlation Coefficient	32
Pearson Correlation Coefficient	22
R-square	2
Area Under the ROC Curve	36
MSE	16
RMSE	17
MAE	18
RSE	20
RRSE	21
RAE	19
Logic Synthesis
Accuracy	5
Error	6
Hits	38
Errors	39
Sensitivity	7
Specificity	8
Positive Predictive Value	9
Negative Predictive Value	10
Recall	33
Precision	34
F1 Measure	35
Jaccard Similarity	31
Matthews Correlation Coefficient	32
Pearson Correlation Coefficient	22
R-square	2
MSE	16
RMSE	17
MAE	18
RSE	20
RRSE	21
RAE	19

See Also:

What's New in GeneXproServer?

Related Tutorials:

How to Create a Mini Cluster with GeneXproServer
Getting Started with Regression
Getting Started with Classification
Getting Started with Logistic Regression
Getting Started with Time Series Prediction
Getting Started with Logic Synthesis

Related Videos:

What's New in GeneXproTools 5.0?

Time Limited Trial

Try GeneXproServer for free for 30 days!

Released February 19, 2014

Last update: 5.0.5667