How to solve "Matrix is rank deficient" Exception

103 views
Skip to first unread message

Achal Saraiya

unread,
Sep 14, 2013, 10:59:57 AM9/14/13
to accor...@googlegroups.com
I am trying to implement Forward variable selection method using R-square criteria. For this I am running Multiple linear regression for multiple time. For this I select random percentage of records. Running this code for multiple time it gives me randomly exception as follow

Additional information: "Matrix is rank deficient"
Source: "Accord.Math"
Trace: " 
at Accord.Math.Decompositions.QrDecomposition.Solve(Double[] value) in D:\Projects\Accord.NET\Assembla\Sources\Accord.Math\Decompositions\QrDecomposition.cs:line 254    
at
Accord.Statistics.Models.Regression.Linear.MultipleLinearRegression.regress(Double[][] inputs, Double[] outputs, Double[,]& designMatrix) in D:\Projects\Accord.NET\Assembla\Sources\Accord.Statistics\Models\Regression\Linear\MultipleLinearRegression.cs:line 256
at
Accord.Statistics.Models.Regression.Linear.MultipleLinearRegression.Regress(Double[][] inputs, Double[] outputs) in D:\Projects\Accord.NET\Assembla\Sources\Accord.Statistics\Models\Regression\Linear\MultipleLinearRegression.cs:line 167    
at
Integrated_Tool_1._4.ForwardVariableSelection.start_regression() in F:\Achal Saraiya\business\freelancing\Varible Selection\Integrated Tool 1.4\Integrated Tool 1.0\VariableSelection.vb:line 446

I don't understand statistic too much. So I don't have Idea if there problem with my code or something related to data. I can I get rid of this problem.

I heartily appreciate for your help

narula.har...@gmail.com

unread,
Jun 16, 2017, 7:20:40 AM6/16/17
to Accord.NET Framework
Were you able to fix this? I have the same problem.
Reply all
Reply to author
Forward
0 new messages