|
0
|
1 <tool id="bg_eden_train" name="EDeN Train" version="0.1">
|
|
|
2 <description></description>
|
|
|
3 <requirements>
|
|
|
4 </requirements>
|
|
|
5 <command>
|
|
|
6 EDeN --action TRAIN
|
|
|
7
|
|
|
8 --input_data_file_name $infile
|
|
|
9 --file_type "SPARSE_VECTOR"
|
|
|
10 --binary_file_type
|
|
|
11
|
|
|
12 ## TODO: we need a tool that creates such a file, maybe from the metadata of an SDF file
|
|
|
13 ## target_file_name is a file with 1 or -1 one in each row, indicating the class
|
|
|
14 --target_file_name $target_infile
|
|
|
15 --model_file_name $model_outfile
|
|
|
16
|
|
|
17 --lambda $lambda ##??? notation?
|
|
|
18 --epochs $epoch
|
|
|
19
|
|
|
20 --sparsification_num_iterations $sparsification_num_iterations
|
|
|
21 --topological_regularization_num_neighbors $topological_regularization_num_neighbors
|
|
|
22 --topological_regularization_decay_rate $topological_regularization_decay_rate
|
|
|
23
|
|
|
24 --num_iterations $num_iterations
|
|
|
25 --threshold $threshold
|
|
|
26 --only_positive $only_positive
|
|
|
27 --only_negative $only_negative
|
|
|
28
|
|
|
29 --random_seed $random_seed
|
|
|
30
|
|
|
31 </command>
|
|
|
32 <inputs>
|
|
|
33 <param format="eden_sparse_vector" name="infile" type="data" label="Input Graph" help=""/>
|
|
|
34 <param format="txt" name="target_infile" type="data" label="Target file" help=""/>
|
|
|
35
|
|
|
36 <param name="kernel_type" type="select" display="radio" label="Type of the Kernel">
|
|
|
37 <option value="NSPDK">NSPDK</option>
|
|
|
38 <option value="WDK">WDK</option>
|
|
|
39 <option value="PBK">PBK</option>
|
|
|
40 <option value="USPK">USPK</option>
|
|
|
41 <option value="DDK">DDK</option>
|
|
|
42 <option value="NSDDK">ANSDDK</option>
|
|
|
43 <option value="SK">SK [NSPDK]</option>
|
|
|
44 </param>
|
|
|
45
|
|
|
46 <param name="graph_type" type="select" display="radio" label="Type of Graph">
|
|
|
47 <option value="DIRECTED">directed</option>
|
|
|
48 <option value="UNDIRECTED">undirected</option>
|
|
|
49 </param>
|
|
|
50
|
|
|
51 <param name="epoch" type="integer" value="10" label="Epoch, Stochastic gradient descend algorithm." help="">
|
|
|
52 <validator type="in_range" min="1" />
|
|
|
53 </param>
|
|
|
54 <param name="lambda" type="text" value="1e-4" label="lambda, Stochastic gradient descend algorithm." help="" />
|
|
|
55
|
|
|
56 </inputs>
|
|
|
57 <outputs>
|
|
|
58 <data format="txt" name="model_outfile" label="Train Model from ${on_string}"/>
|
|
|
59 </outputs>
|
|
|
60 <tests>
|
|
|
61 <test>
|
|
|
62 <param name="infile" value="3_molceuls.sdf" />
|
|
|
63 <output name="outfile" file="3_molecules.gspan" />
|
|
|
64 </test>
|
|
|
65 </tests>
|
|
|
66 <help>
|
|
|
67
|
|
|
68 .. class:: infomark
|
|
|
69
|
|
|
70 **What it does**
|
|
|
71
|
|
|
72 The linear model is induced using the accelerated stochastic gradient descent technique by Léon Bottou and Yann LeCun.
|
|
|
73 When the target information is 0, a self-training algorithm is used to impute a positive or negative class to the unsupervised instances.
|
|
|
74 If the target information is imbalanced a minority class resampling technique is used to rebalance the training set.
|
|
|
75
|
|
|
76 This tool is part of the EDeN (Explicit Decomposition with Neighborhoods) suite, developed by Fabrizio Costa.
|
|
|
77
|
|
|
78
|
|
|
79 REFERENCES
|
|
|
80 ==========
|
|
|
81
|
|
|
82 The code for Stochastic Gradient Descent SVM is adapted from http://leon.bottou.org/projects/sgd. Léon Bottou and Yann LeCun, ''Large Scale Online Learning'', Advances in Neural Information Processing Systems 16, Edited by Sebastian Thrun, Lawrence Saul and Bernhard Schölkopf, MIT Press, Cambridge, MA, 2004.
|
|
|
83
|
|
|
84
|
|
|
85
|
|
|
86 </help>
|
|
|
87 </tool>
|
