[jira] [Commented] (MADLIB-1222) Support already encoded arrays for dependent var in MLP classification
[ https://issues.apache.org/jira/browse/MADLIB-1222?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16426151#comment-16426151 ] ASF GitHub Bot commented on MADLIB-1222: Github user asfgit closed the pull request at: https://github.com/apache/madlib/pull/250 > Support already encoded arrays for dependent var in MLP classification > -- > > Key: MADLIB-1222 > URL: https://issues.apache.org/jira/browse/MADLIB-1222 > Project: Apache MADlib > Issue Type: New Feature > Components: Module: Neural Networks >Reporter: Nandish Jayaram >Priority: Major > Fix For: v1.14 > > > MLP currently only supports scalar dependent variables for MLP > classification. If a user has already one-hot encoded categorical variables > the dependent variable will be an array, and hence unusable with > mlp_classification. This feature request is to allow the use of one-hot > encoded array for dependent vars in MLP classification. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (MADLIB-1222) Support already encoded arrays for dependent var in MLP classification
[ https://issues.apache.org/jira/browse/MADLIB-1222?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16426125#comment-16426125 ] ASF GitHub Bot commented on MADLIB-1222: Github user fmcquillan99 commented on the issue: https://github.com/apache/madlib/pull/250 See JIRA https://issues.apache.org/jira/browse/MADLIB-1222 for examples showing this works for IGD and mini-batch LGTM I think u can go ahead and merge this PR to master > Support already encoded arrays for dependent var in MLP classification > -- > > Key: MADLIB-1222 > URL: https://issues.apache.org/jira/browse/MADLIB-1222 > Project: Apache MADlib > Issue Type: New Feature > Components: Module: Neural Networks >Reporter: Nandish Jayaram >Priority: Major > Fix For: v1.14 > > > MLP currently only supports scalar dependent variables for MLP > classification. If a user has already one-hot encoded categorical variables > the dependent variable will be an array, and hence unusable with > mlp_classification. This feature request is to allow the use of one-hot > encoded array for dependent vars in MLP classification. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (MADLIB-1222) Support already encoded arrays for dependent var in MLP classification
[
https://issues.apache.org/jira/browse/MADLIB-1222?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16426112#comment-16426112
]
Frank McQuillan commented on MADLIB-1222:
-
For minibatch this seems to work OK. e.g., continuing modified version user
docs example from above:
{code:sql}
DROP TABLE IF EXISTS iris_data_packed, iris_data_packed_standardization,
iris_data_packed_summary;
SELECT madlib.minibatch_preprocessor(
'iris_data',
'iris_data_packed',
'class_integer',
'attributes',
10
);
{code}
{code:sql}
DROP TABLE IF EXISTS mlp_model, mlp_model_summary, mlp_model_standardization;
-- Set seed so results are reproducible
SELECT setseed(0);
SELECT madlib.mlp_classification(
'iris_data_packed', -- Source table
'mlp_model', -- Destination table
'independent_varname', -- Input features
'dependent_varname', -- Label
ARRAY[5], -- Number of units per layer
'learning_rate_init=0.003,
n_iterations=500,
tolerance=0', -- Optimizer params
'tanh', -- Activation function
NULL, -- Default weight (1)
FALSE,-- No warm start
FALSE -- Not verbose
);
{code}
{code:sql}
DROP TABLE IF EXISTS mlp_prediction;
SELECT madlib.mlp_predict(
'mlp_model', -- Model table
'iris_data', -- Test data table
'id',-- Id column in test table
'mlp_prediction',-- Output table for predictions
'response' -- Output classes, not probabilities
);
SELECT * FROM mlp_prediction JOIN iris_data USING (id) ORDER BY id;
{code}
produces
{code}
id | estimated_class_integer |attributes | class_integer | class |
state
+-+---+---+---+---
1 | {1,0} | {5.0,3.2,1.2,0.2} | {1,0} | 1 |
Alaska
2 | {1,0} | {5.5,3.5,1.3,0.2} | {1,0} | 1 |
Alaska
3 | {1,0} | {4.9,3.1,1.5,0.1} | {1,0} | 1 |
Alaska
4 | {1,0} | {4.4,3.0,1.3,0.2} | {1,0} | 1 |
Alaska
5 | {1,0} | {5.1,3.4,1.5,0.2} | {1,0} | 1 |
Alaska
6 | {1,0} | {5.0,3.5,1.3,0.3} | {1,0} | 1 |
Alaska
7 | {1,0} | {4.5,2.3,1.3,0.3} | {1,0} | 1 |
Alaska
8 | {1,0} | {4.4,3.2,1.3,0.2} | {1,0} | 1 |
Alaska
9 | {1,0} | {5.0,3.5,1.6,0.6} | {1,0} | 1 |
Alaska
10 | {1,0} | {5.1,3.8,1.9,0.4} | {1,0} | 1 |
Alaska
11 | {1,0} | {4.8,3.0,1.4,0.3} | {1,0} | 1 |
Alaska
12 | {1,0} | {5.1,3.8,1.6,0.2} | {1,0} | 1 |
Alaska
13 | {0,1} | {5.7,2.8,4.5,1.3} | {0,1} | 2 |
Alaska
14 | {0,1} | {6.3,3.3,4.7,1.6} | {0,1} | 2 |
Alaska
15 | {0,1} | {4.9,2.4,3.3,1.0} | {0,1} | 2 |
Alaska
16 | {0,1} | {6.6,2.9,4.6,1.3} | {0,1} | 2 |
Alaska
17 | {0,1} | {5.2,2.7,3.9,1.4} | {0,1} | 2 |
Alaska
18 | {0,1} | {5.0,2.0,3.5,1.0} | {0,1} | 2 |
Alaska
19 | {0,1} | {5.9,3.0,4.2,1.5} | {0,1} | 2 |
Alaska
20 | {0,1} | {6.0,2.2,4.0,1.0} | {0,1} | 2 |
Alaska
21 | {0,1} | {6.1,2.9,4.7,1.4} | {0,1} | 2 |
Alaska
22 | {0,1} | {5.6,2.9,3.6,1.3} | {0,1} | 2 |
Alaska
23 | {0,1} | {6.7,3.1,4.4,1.4} | {0,1} | 2 |
Alaska
24 | {0,1} | {5.6,3.0,4.5,1.5} | {0,1} | 2 |
Alaska
25 | {0,1} | {5.8,2.7,4.1,1.0} | {0,1} | 2 |
Alaska
26 | {0,1} | {6.2,2.2,4.5,1.5} | {0,1} | 2 |
Alaska
27 | {0,1} | {5.6,2.5,3.9,1.1} | {0,1} | 2 |
Alaska
28 | {1,0} | {5.0,3.4,1.5,0.2} | {1,0} | 1 |
Tennessee
29 | {1,0} | {4.4,2.9,1.4,0.2} | {1,0} | 1 |
Tennessee
30 | {1,0} | {4.9,3.1,1.5,0.1} | {1,0} | 1 |
Tennessee
31 | {1,0} | {5.4,3.7,1.5,0.2} | {1,0} | 1 |
Tennessee
32 | {1,0} | {4.8,3.4,1.6,0.2} | {1,0} | 1 |
Tennessee
33 | {1,0} | {4.8,3.0,1.4,0.1} | {1,0} | 1 |
Tennessee
34 | {1,0} | {4.3,3.0,1.1,0.1} | {1,0} | 1 |
Tennessee
35 | {1,0} | {5.8,4.0,1.2,0.2} | {1,0} | 1 |
Tennessee
36 | {1,0} | {5.7,4.4,1.5,0.4} | {1,0} | 1 |
Tennes
[jira] [Commented] (MADLIB-1222) Support already encoded arrays for dependent var in MLP classification
[
https://issues.apache.org/jira/browse/MADLIB-1222?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16426074#comment-16426074
]
Frank McQuillan commented on MADLIB-1222:
-
For IGD (without minibatch) this seems to work OK. e.g., using a modified
version user docs example:
[http://madlib.apache.org/docs/latest/group__grp__nn.html#example]
{code:sql}
DROP TABLE IF EXISTS iris_data;
CREATE TABLE iris_data(
id serial,
attributes numeric[],
class_integer integer[],
class integer,
state varchar
);
INSERT INTO iris_data(id, attributes, class_integer, class, state) VALUES
(1,ARRAY[5.0,3.2,1.2,0.2], ARRAY[1,0],1,'Alaska'),
(2,ARRAY[5.5,3.5,1.3,0.2], ARRAY[1,0],1,'Alaska'),
(3,ARRAY[4.9,3.1,1.5,0.1], ARRAY[1,0],1,'Alaska'),
(4,ARRAY[4.4,3.0,1.3,0.2], ARRAY[1,0],1,'Alaska'),
(5,ARRAY[5.1,3.4,1.5,0.2], ARRAY[1,0],1,'Alaska'),
(6,ARRAY[5.0,3.5,1.3,0.3], ARRAY[1,0],1,'Alaska'),
(7,ARRAY[4.5,2.3,1.3,0.3], ARRAY[1,0],1,'Alaska'),
(8,ARRAY[4.4,3.2,1.3,0.2], ARRAY[1,0],1,'Alaska'),
(9,ARRAY[5.0,3.5,1.6,0.6], ARRAY[1,0],1,'Alaska'),
(10,ARRAY[5.1,3.8,1.9,0.4], ARRAY[1,0],1,'Alaska'),
(11,ARRAY[4.8,3.0,1.4,0.3], ARRAY[1,0],1,'Alaska'),
(12,ARRAY[5.1,3.8,1.6,0.2], ARRAY[1,0],1,'Alaska'),
(13,ARRAY[5.7,2.8,4.5,1.3], ARRAY[0,1],2,'Alaska'),
(14,ARRAY[6.3,3.3,4.7,1.6], ARRAY[0,1],2,'Alaska'),
(15,ARRAY[4.9,2.4,3.3,1.0], ARRAY[0,1],2,'Alaska'),
(16,ARRAY[6.6,2.9,4.6,1.3], ARRAY[0,1],2,'Alaska'),
(17,ARRAY[5.2,2.7,3.9,1.4], ARRAY[0,1],2,'Alaska'),
(18,ARRAY[5.0,2.0,3.5,1.0], ARRAY[0,1],2,'Alaska'),
(19,ARRAY[5.9,3.0,4.2,1.5], ARRAY[0,1],2,'Alaska'),
(20,ARRAY[6.0,2.2,4.0,1.0], ARRAY[0,1],2,'Alaska'),
(21,ARRAY[6.1,2.9,4.7,1.4], ARRAY[0,1],2,'Alaska'),
(22,ARRAY[5.6,2.9,3.6,1.3], ARRAY[0,1],2,'Alaska'),
(23,ARRAY[6.7,3.1,4.4,1.4], ARRAY[0,1],2,'Alaska'),
(24,ARRAY[5.6,3.0,4.5,1.5], ARRAY[0,1],2,'Alaska'),
(25,ARRAY[5.8,2.7,4.1,1.0], ARRAY[0,1],2,'Alaska'),
(26,ARRAY[6.2,2.2,4.5,1.5], ARRAY[0,1],2,'Alaska'),
(27,ARRAY[5.6,2.5,3.9,1.1], ARRAY[0,1],2,'Alaska'),
(28,ARRAY[5.0,3.4,1.5,0.2], ARRAY[1,0],1,'Tennessee'),
(29,ARRAY[4.4,2.9,1.4,0.2], ARRAY[1,0],1,'Tennessee'),
(30,ARRAY[4.9,3.1,1.5,0.1], ARRAY[1,0],1,'Tennessee'),
(31,ARRAY[5.4,3.7,1.5,0.2], ARRAY[1,0],1,'Tennessee'),
(32,ARRAY[4.8,3.4,1.6,0.2], ARRAY[1,0],1,'Tennessee'),
(33,ARRAY[4.8,3.0,1.4,0.1], ARRAY[1,0],1,'Tennessee'),
(34,ARRAY[4.3,3.0,1.1,0.1], ARRAY[1,0],1,'Tennessee'),
(35,ARRAY[5.8,4.0,1.2,0.2], ARRAY[1,0],1,'Tennessee'),
(36,ARRAY[5.7,4.4,1.5,0.4], ARRAY[1,0],1,'Tennessee'),
(37,ARRAY[5.4,3.9,1.3,0.4], ARRAY[1,0],1,'Tennessee'),
(38,ARRAY[6.0,2.9,4.5,1.5], ARRAY[0,1],2,'Tennessee'),
(39,ARRAY[5.7,2.6,3.5,1.0], ARRAY[0,1],2,'Tennessee'),
(40,ARRAY[5.5,2.4,3.8,1.1], ARRAY[0,1],2,'Tennessee'),
(41,ARRAY[5.5,2.4,3.7,1.0], ARRAY[0,1],2,'Tennessee'),
(42,ARRAY[5.8,2.7,3.9,1.2], ARRAY[0,1],2,'Tennessee'),
(43,ARRAY[6.0,2.7,5.1,1.6], ARRAY[0,1],2,'Tennessee'),
(44,ARRAY[5.4,3.0,4.5,1.5], ARRAY[0,1],2,'Tennessee'),
(45,ARRAY[6.0,3.4,4.5,1.6], ARRAY[0,1],2,'Tennessee'),
(46,ARRAY[6.7,3.1,4.7,1.5], ARRAY[0,1],2,'Tennessee'),
(47,ARRAY[6.3,2.3,4.4,1.3], ARRAY[0,1],2,'Tennessee'),
(48,ARRAY[5.6,3.0,4.1,1.3], ARRAY[0,1],2,'Tennessee'),
(49,ARRAY[5.5,2.5,4.0,1.3], ARRAY[0,1],2,'Tennessee'),
(50,ARRAY[5.5,2.6,4.4,1.2], ARRAY[0,1],2,'Tennessee'),
(51,ARRAY[6.1,3.0,4.6,1.4], ARRAY[0,1],2,'Tennessee'),
(52,ARRAY[5.8,2.6,4.0,1.2], ARRAY[0,1],2,'Tennessee');
{code}
{code:sql}
DROP TABLE IF EXISTS mlp_model, mlp_model_summary, mlp_model_standardization;
-- Set seed so results are reproducible
SELECT setseed(0);
SELECT madlib.mlp_classification(
'iris_data', -- Source table
'mlp_model', -- Destination table
'attributes', -- Input features
'class_integer', -- Label
ARRAY[5], -- Number of units per layer
'learning_rate_init=0.003,
n_iterations=500,
tolerance=0', -- Optimizer params
'tanh', -- Activation function
NULL, -- Default weight (1)
FALSE,-- No warm start
FALSE -- Not verbose
);
{code}
{code:sql}
DROP TABLE IF EXISTS mlp_prediction;
SELECT madlib.mlp_predict(
'mlp_model', -- Model table
'iris_data', -- Test data table
'id',-- Id column in test table
'mlp_prediction',-- Output table for predictions
'prob' -- Output classes, not probabilities
);
SELECT * FROM mlp_prediction JOIN iris_data USING (id) ORDER BY id;
{code}
produces
{code}
id | estimated_class_integer |attributes | class_integer | class |
state
+-+---+---+---+---
1 | {1,0} | {5.0,3.2,1.2,0.2} | {1,0} | 1 |
Alaska
2 | {1,0} | {5.5,3.5,1.3,0.2} | {1,0} | 1 |
Alaska
3 | {1,0} | {4.9,3.1,1.5,0.1} | {1,0} | 1 |
Alaska
[jira] [Commented] (MADLIB-1222) Support already encoded arrays for dependent var in MLP classification
[ https://issues.apache.org/jira/browse/MADLIB-1222?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16414347#comment-16414347 ] ASF GitHub Bot commented on MADLIB-1222: GitHub user njayaram2 opened a pull request: https://github.com/apache/madlib/pull/250 MLP: Allow one-hot encoded dependent var for classification JIRA:MADLIB-1222 MLP currently automatically encodes categorical variables for classification but does not allow already encoded arrays for dependent variables in mlp_classification. This commit lets users have an already encoded array for the dependent variable and train a model. You can merge this pull request into a Git repository by running: $ git pull https://github.com/madlib/madlib feature/mlp/support-encoded-dep-var Alternatively you can review and apply these changes as the patch at: https://github.com/apache/madlib/pull/250.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #250 commit f5a87dee6bc8f27c1a13a4921ea726b391b1813d Author: Nandish Jayaram Date: 2018-03-20T22:43:25Z MLP: Allow one-hot encoded dependent var for classification JIRA:MADLIB-1222 MLP currently automatically encodes categorical variables for classification but does not allow already encoded arrays for dependent variables in mlp_classification. This commit lets users have an already encoded array for the dependent variable and train a model. > Support already encoded arrays for dependent var in MLP classification > -- > > Key: MADLIB-1222 > URL: https://issues.apache.org/jira/browse/MADLIB-1222 > Project: Apache MADlib > Issue Type: New Feature > Components: Module: Neural Networks >Reporter: Nandish Jayaram >Priority: Major > Fix For: v1.14 > > > MLP currently only supports scalar dependent variables for MLP > classification. If a user has already one-hot encoded categorical variables > the dependent variable will be an array, and hence unusable with > mlp_classification. This feature request is to allow the use of one-hot > encoded array for dependent vars in MLP classification. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (MADLIB-1222) Support already encoded arrays for dependent var in MLP classification
[
https://issues.apache.org/jira/browse/MADLIB-1222?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16412215#comment-16412215
]
Nandish Jayaram commented on MADLIB-1222:
-
Example use case, and handling it:
1) User encodes dep vars with whatever tool they want and put it in the column
`color` . Maybe they do this to anonymize, maybe the data is just in that
format already:
{code}
blue [1,0,0]
red [0,1,0]
green [0,0,1]
{code}
---
start MADlib
2) runs mini-batch preprocess (if planning to use mini-batch)
3) runs MLP classification train (IGD or mini-batch)
4a) runs MLP predict (response):
{code}
actual predicted
[0,1,0] [1,0,0]
[0,0,1] [0,0,1]
[1,0,0] [1,0,0]
etc.
{code}
4b) runs MLP predict (prob):
{code}
actual estimated_prob
[0,1,0] [0.85, 0.10, ,0.05]
[0,0,1] [0.0 , 0.1 , 0.9]
[1,0,0] [0.75, 0.20, 0.05]
etc.
{code}
end MADlib
--
5) User maps back to red, blue, green since they know the mapping but MADlib
doesn't.
> Support already encoded arrays for dependent var in MLP classification
> --
>
> Key: MADLIB-1222
> URL: https://issues.apache.org/jira/browse/MADLIB-1222
> Project: Apache MADlib
> Issue Type: New Feature
> Components: Module: Neural Networks
>Reporter: Nandish Jayaram
>Priority: Major
> Fix For: v1.14
>
>
> MLP currently only supports scalar dependent variables for MLP
> classification. If a user has already one-hot encoded categorical variables
> the dependent variable will be an array, and hence unusable with
> mlp_classification. This feature request is to allow the use of one-hot
> encoded array for dependent vars in MLP classification.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
