ENM SVM workflow used for optimization

Created: 2013-08-09 16:01:50 Last updated: 2013-09-04 07:20:28

Download Workflow

This workflow was used for the optimization of the SVM algorithm from the openModeller toolbox (http://openmodeller.sourceforge.net). The workflow uses 10-fold cross-validation and then calculates the average AUC, which can be used as fitness value during parameter optimization.

Preview

Download as scalable diagram (SVG)

Run

Run this Workflow in the Taverna Workbench...

Option 1:

Copy and paste this link into File > 'Open workflow location...'
http://www.myexperiment.org/workflows/3680/download?version=1
[ More Info Expand ]

Workflow Components

Authors (1)

Titles (1)

Descriptions (1)

Dependencies (0)

Inputs (11)

Name	Description
input_points	This input takes a text file containing species occurrence points. The file must be formatted according to the Darwin Core archive standard: each line corresponds to a different record, values are separated by comma, and the first line must be a header (also separated by comma) containing the Darwin Core terms for each column. The following terms are mandatory to run this workflow: occurrenceID, nameComplete, decimalLongitude and decimalLatitude. Only records for the first species will be used.
SvmType	Type of SVM: 0 = C-SVC, 1 = Nu-SVC, 2 = one-class SVM
KernelType	Type of kernel function: 0 = linear: u'v , 1 = polynomial: (gammau'v + coef0)^degree , 2 = radial basis function: exp(-gamma\|u-v\|^2)
Degree	Degree in kernel function (only for polynomial kernels). Domain: [0, oo)
Gamma	Gamma in kernel function (only for polynomial and radial basis kernels). When set to zero, the default value will actually be 1/k, where k is the number of layers. Domain: (oo, oo)
Coef0	Coef0 in kernel function (only for polynomial kernels). Domain: (oo, oo).
C	Cost (only for C-SVC types). Domain: (0.0, oo)
Nu	Nu (only for Nu-SVC). Domain: (0.0, 1.0]
ProbabilisticOutput	Indicates if the output should be a probability instead of a binary response (only available for C-SVC and Nu-SVC). Possible values: 0 or 1.
NumberOfPseudoAbsences	Number of pseudo absences to be generated when the SVM type is C-SVC or Nu-SVC. Zero means the same number of presence points.
mask

Processors (8)

Name	Type	Description
select_algorithm	workflow
parse_input_points	workflow
select_layers	workflow
allocate_points	workflow
create_and_test_model	workflow
extract_aucs	workflow
calculate_mean	workflow
Flatten_List	localworker	Script flatten(inputs, outputs, depth) { for (i = inputs.iterator(); i.hasNext();) { element = i.next(); if (element instanceof Collection && depth > 0) { flatten(element, outputs, depth - 1); } else { outputs.add(element); } } } outputlist = new ArrayList(); flatten(inputlist, outputlist, 1);

Beanshells (6)

Name	Inputs	Outputs
allocate_points	csv_points id_idx long_idx lat_idx folds	training_points testing_points all_points num_points
create_algorithm_xml	algorithm_id algorithm_version svm_type kernel_type degree gamma coef0 c nu probabilistic_output number_of_pseudo_absences	createmodel_algorithm
calculate_mean	auc_list	mean_auc
parse_header	csv_content	name_idx id_idx long_idx lat_idx
get_first_taxon	csv_content name_idx	taxon_name
extract_taxon_points	csv_content name_idx taxon_name	taxon_points

Outputs (1)

Name	Description
auc

Datalinks (23)

Source	Sink
SvmType	select_algorithm:SvmType
KernelType	select_algorithm:KernelType
Degree	select_algorithm:Degree
Gamma	select_algorithm:Gamma
Coef0	select_algorithm:Coef0
C	select_algorithm:C
Nu	select_algorithm:Nu
ProbabilisticOutput	select_algorithm:ProbabilisticOutput
NumberOfPseudoAbsences	select_algorithm:NumberOfPseudoAbsences
input_points	parse_input_points:csv_content
parse_input_points:id_idx	allocate_points:id_idx
parse_input_points:lat_idx	allocate_points:lat_idx
parse_input_points:long_idx	allocate_points:long_idx
parse_input_points:taxon_points	allocate_points:csv_points
select_algorithm:algorithm	create_and_test_model:algorithm_xml
allocate_points:testing_points	create_and_test_model:testing_points
select_layers:selected_layers	create_and_test_model:layers_str
allocate_points:training_points	create_and_test_model:training_points
mask	create_and_test_model:mask
create_and_test_model:test_model_statistics	extract_aucs:test_statistics_xml
Flatten_List:outputlist	calculate_mean:auc_list
extract_aucs:auc	Flatten_List:inputlist
calculate_mean:mean_auc	auc

Coordinations (1)

Controller	Target
create_and_test_model	extract_aucs

Information Workflow Type

Taverna 2

Information Uploader

Sonja Holl

Information License

All versions of this Workflow are licensed under:

Information Version 1 (of 1)

Information Credits (2)

(People/Groups)

Information Attributions (0)

(Workflows/Files)

None

Information Tags (6)

Uploader tags

auc
|
biovel
|
enm
|
optimization
|
svm
|
workflow

Log in to add Tags

Information Shared with Groups (0)

None

Information Featured In Packs (0)

None

Log in to add to one of your Packs

Information Attributed By (0)

(Workflows/Files)

None

Information Favourited By (0)

No one

Information Statistics

2409 viewings

1214 downloads

[ see breakdown ]

Citations (0)

None

Version History

In chronological order:

ENM SVM workflow used for optimization

Created by Sonja Holl on Friday 09 August 2013 16:01:49 (UTC)

Last edited by Sonja Holl on Wednesday 04 September 2013 07:20:27 (UTC)

Reviews (0)

No reviews yet

Be the first to review!

Comments (0)

View Timeline

No comments yet

Log in to make a comment

Other workflows that use similar services (2)

Taverna 2

Uploader

Aleksandra Pawlik

BioVeL workshop reduced full workflow (1)

Download

The workflow is a reduced version of the main workflow. It can be used to call web services to support biodiversity research. It fetches data from GBIF. The occurrences are then used to create a model using OpenModeller, test the model and to project the model. The projection of the model is currently a "native projection" i.e. it uses the same layers as those used to create the model. The services within the workflow will allow non-native projections

Created: 2013-09-04

Credits: Katy Wolstencroft Alan Williams

Attributions:

Taverna 2

Uploader

Sonja Holl

ENM Maxent workflow used for optimization (1)

Download

This workflow was used for the optimization of the Maxent algorithm from the openModeller toolbox (http://openmodeller.sourceforge.net/). The workflow uses 10-fold cross-validation and then calculates the average AUC, which can be used as fitness value during parameter optimization.

Created: 2013-08-09 | Last updated: 2013-09-04

Credits: Sonja Holl Renato De Giovanni

ENM SVM workflow used for optimization

Preview

Run

Run this Workflow in the Taverna Workbench...

Workflow Components

Script

Reviews (0)

Comments (0)

Other workflows that use similar services (2)