Mercurial > repos > bgruening > svm_classifier

comparison svm.xml @ 0:9a9396e5d153 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/bgruening/galaxytools/tools/sklearn commit 0e582cf1f3134c777cce3aa57d71b80ed95e6ba9
author bgruening
date Fri, 16 Feb 2018 09:16:30 -0500
parents
children 78c664cc1841
comparison
equal deleted inserted replaced
-1:000000000000 0:9a9396e5d153
1 <tool id="svm_classifier" name="Support vector machines (SVMs)" version="@VERSION@">
2 <description>for classification</description>
3 <macros>
4 <import>main_macros.xml</import>
5 <!-- macro name="class_weight" argument="class_weight"-->
6 </macros>
7 <expand macro="python_requirements"/>
8 <expand macro="macro_stdio"/>
9 <version_command>echo "@VERSION@"</version_command>
10 <command><![CDATA[
11 python "$svc_script" '$inputs'
12 ]]>
13 </command>
14 <configfiles>
15 <inputs name="inputs"/>
16 <configfile name="svc_script">
17 <![CDATA[
18 import sys
19 import json
20 import numpy as np
21 import sklearn.svm
22 import pandas
23 import pickle
24
25 input_json_path = sys.argv[1]
26 params = json.load(open(input_json_path, "r"))
27
28 #if $selected_tasks.selected_task == "load":
29
30 classifier_object = pickle.load(open("$infile_model", 'rb'))
31
32 data = pandas.read_csv("$selected_tasks.infile_data", sep='\t', header=0, index_col=None, parse_dates=True, encoding=None, tupleize_cols=False )
33 prediction = classifier_object.predict(data)
34 prediction_df = pandas.DataFrame(prediction)
35 res = pandas.concat([data, prediction_df], axis=1)
36 res.to_csv(path_or_buf = "$outfile_predict", sep="\t", index=False)
37
38 #else:
39
40 data_train = pandas.read_csv("$selected_tasks.infile_train", sep='\t', header=0, index_col=None, parse_dates=True, encoding=None, tupleize_cols=False )
41
42 data = data_train.ix[:,0:len(data_train.columns)-1]
43 labels = np.array(data_train[data_train.columns[len(data_train.columns)-1]])
44
45 options = params["selected_tasks"]["selected_algorithms"]["options"]
46 selected_algorithm = params["selected_tasks"]["selected_algorithms"]["selected_algorithm"]
47
48 if not(selected_algorithm=="LinearSVC"):
49 if options["kernel"]:
50 options["kernel"] = str(options["kernel"])
51
52 my_class = getattr(sklearn.svm, selected_algorithm)
53 classifier_object = my_class(**options)
54 classifier_object.fit(data,labels)
55
56 pickle.dump(classifier_object,open("$outfile_fit", 'w+'))
57
58 #end if
59
60 ]]>
61 </configfile>
62 </configfiles>
63 <inputs>
64 <expand macro="train_loadConditional" model="zip">
65 <param name="selected_algorithm" type="select" label="Classifier type">
66 <option value="SVC">C-Support Vector Classification</option>
67 <option value="NuSVC">Nu-Support Vector Classification</option>
68 <option value="LinearSVC">Linear Support Vector Classification</option>
69 </param>
70 <when value="SVC">
71 <expand macro="svc_advanced_options">
72 <expand macro="C"/>
73 </expand>
74 </when>
75 <when value="NuSVC">
76 <expand macro="svc_advanced_options">
77 <param argument="nu" type="float" optional="true" value="0.5" label="Nu control parameter" help="Controls the number of support vectors. Should be in the interval (0, 1]. "/>
78 </expand>
79 </when>
80 <when value="LinearSVC">
81 <section name="options" title="Advanced Options" expanded="False">
82 <expand macro="C"/>
83 <expand macro="tol" default_value="0.001" help_text="Tolerance for stopping criterion. "/>
84 <expand macro="random_state" help_text="Integer number. The seed of the pseudo random number generator to use when shuffling the data for probability estimation. A fixed seed allows reproducible results."/>
85 <!--expand macro="class_weight"/-->
86 <param argument="max_iter" type="integer" optional="true" value="1000" label="Maximum number of iterations" help="The maximum number of iterations to be run."/>
87 <param argument="loss" type="select" label="Loss function" help="Specifies the loss function. ''squared_hinge'' is the square of the hinge loss.">
88 <option value="squared_hinge" selected="true">Squared hinge</option>
89 <option value="hinge">Hinge</option>
90 </param>
91 <param argument="penalty" type="select" label="Penalization norm" help=" ">
92 <option value="l1" >l1</option>
93 <option value="l2" selected="true">l2</option>
94 </param>
95 <param argument="dual" type="boolean" optional="true" truevalue="booltrue" falsevalue="boolflase" checked="true" label="Use the shrinking heuristic" help="Select the algorithm to either solve the dual or primal optimization problem. Prefer dual=False when n_samples > n_features."/>
96 <param argument="multi_class" type="select" label="Multi-class strategy" help="Determines the multi-class strategy if y contains more than two classes.">
97 <option value="ovr" selected="true">ovr</option>
98 <option value="crammer_singer" >crammer_singer</option>
99 </param>
100 <param argument="fit_intercept" type="boolean" optional="true" truevalue="booltrue" falsevalue="boolflase" checked="true" label="Calculate the intercept for this model" help="If set to false, data is expected to be already centered."/>
101 <param argument="intercept_scaling" type="float" optional="true" value="1" label="Add synthetic feature to the instance vector" help=" "/>
102 </section>
103 </when>
104 </expand>
105 </inputs>
106
107 <expand macro="output"/>
108
109 <tests>
110 <test>
111 <param name="infile_train" value="train_set.tabular" ftype="tabular"/>
112 <param name="selected_task" value="train"/>
113 <param name="selected_algorithm" value="SVC"/>
114 <param name="random_state" value="5"/>
115 <output name="outfile_fit" file="svc_model01.txt"/>
116 </test>
117 <test>
118 <param name="infile_train" value="train_set.tabular" ftype="tabular"/>
119 <param name="selected_task" value="train"/>
120 <param name="selected_algorithm" value="NuSVC"/>
121 <param name="random_state" value="5"/>
122 <output name="outfile_fit" file="svc_model02.txt"/>
123 </test>
124 <test>
125 <param name="infile_train" value="train_set.tabular" ftype="tabular"/>
126 <param name="selected_task" value="train"/>
127 <param name="selected_algorithm" value="LinearSVC"/>
128 <param name="random_state" value="5"/>
129 <output name="outfile_fit" file="svc_model03.txt"/>
130 </test>
131 <test>
132 <param name="infile_model" value="svc_model01.txt" ftype="txt"/>
133 <param name="infile_data" value="test_set.tabular" ftype="tabular"/>
134 <param name="selected_task" value="load"/>
135 <output name="outfile_predict" file="svc_prediction_result01.tabular"/>
136 </test>
137 <test>
138 <param name="infile_model" value="svc_model02.txt" ftype="txt"/>
139 <param name="infile_data" value="test_set.tabular" ftype="tabular"/>
140 <param name="selected_task" value="load"/>
141 <output name="outfile_predict" file="svc_prediction_result02.tabular"/>
142 </test>
143 <test>
144 <param name="infile_model" value="svc_model03.txt" ftype="txt"/>
145 <param name="infile_data" value="test_set.tabular" ftype="tabular"/>
146 <param name="selected_task" value="load"/>
147 <output name="outfile_predict" file="svc_prediction_result03.tabular"/>
148 </test>
149 </tests>
150 <help><![CDATA[
151 **What it does**
152 This module implements the Support Vector Machine (SVM) classification algorithms.
153 Support vector machines (SVMs) are a set of supervised learning methods used for classification, regression and outliers detection.
154
155 **The advantages of support vector machines are:**
156
157 1- Effective in high dimensional spaces.
158
159 2- Still effective in cases where number of dimensions is greater than the number of samples.
160
161 3- Uses a subset of training points in the decision function (called support vectors), so it is also memory efficient.
162
163 4- Versatile: different Kernel functions can be specified for the decision function. Common kernels are provided, but it is also possible to specify custom kernels.
164
165 **The disadvantages of support vector machines include:**
166
167 1- If the number of features is much greater than the number of samples, the method is likely to give poor performances.
168
169 2- SVMs do not directly provide probability estimates, these are calculated using an expensive five-fold cross-validation
170
171 For more information check http://scikit-learn.org/stable/modules/neighbors.html
172
173 ]]>
174 </help>
175 <expand macro="sklearn_citation"/>
176 </tool>