python - sklearn.linear_model.LogisticRegression returns different coefficients every time although random_state is set -

-

i'm fitting logistic regression model , setting random state fixed value.

every time "fit" different coefficients, example:

classifier_instance.fit(train_examples_features, train_examples_labels) logisticregression(c=1.0, class_weight=none, dual=false, fit_intercept=true,           intercept_scaling=1, penalty='l2', random_state=1, tol=0.0001)  >>> classifier_instance.raw_coef_ array([[ 0.071101940040772596  ,  0.05143724979709707323,  0.071101940040772596  , -0.04089477198935181912, -0.0407380696457252528 ,  0.03622160087086594843,  0.01055345545606742319,          0.01071861708285645406, -0.36248634699444892693, -0.06159019047096317423,  0.02370064668025737009,  0.02370064668025737009, -0.03159781822495803805,  0.11221150783553821006,          0.02728295348681779309,  0.071101940040772596  ,  0.071101940040772596  ,  0.                    ,  0.10882033432637286396,  0.64630314505709030026,  0.09617956519989406816,          0.0604133873444507169 ,  0.                    ,  0.04111685986987245051,  0.                    ,  0.                    ,  0.18312324521915510078,  0.071101940040772596  ,          0.071101940040772596  ,  0.                    , -0.59561802045324663268, -0.61490898457874587635,  1.07812569991461248975,  0.071101940040772596  ]])  classifier_instance.fit(train_examples_features, train_examples_labels) logisticregression(c=1.0, class_weight=none, dual=false, fit_intercept=true,           intercept_scaling=1, penalty='l2', random_state=1, tol=0.0001)  >>> classifier_instance.raw_coef_ array([[ 0.07110193825129411394,  0.05143724970282205489,  0.07110193825129411394, -0.04089477178162870957, -0.04073806899140903354,  0.03622160048165772028,  0.010553455400928528  ,          0.01071860364222424096, -0.36248635488413910588, -0.06159021545062405567,  0.02370064608376460866,  0.02370064608376460866, -0.03159783710841745225,  0.11221149816037970237,          0.02728295411479400578,  0.07110193825129411394,  0.07110193825129411394,  0.                    ,  0.10882033461822394893,  0.64630314701686075729,  0.09617956493834901865,          0.06041338563697066372,  0.                    ,  0.04111676713793514099,  0.                    ,  0.                    ,  0.18312324401049043243,  0.07110193825129411394,          0.07110193825129411394,  0.                    , -0.59561803345113684127, -0.61490899867901249731,  1.07812569539027203191,  0.07110193825129411394]])

i'm using version 0.14, docs specify "the underlying c implementation uses random number generator select features when fitting model. not uncommon, have different results same input data. if happens, try smaller tol parameter."

i thought setting random state make sure there no randomness apparently not case. bug or desired behavior?

it's not desired, it's known issue hard fix. thing logisticregression models trained liblinear, not allow setting random seed in robust way. when explicitly set random_state, best effort made set liblinear's random seed, may fail.


Comments

Post a Comment

Popular posts from this blog

java - How to print docx and pdf? -

-
i can print using txt file printer, however, when change docx or pdf many undefined characters cannot understood @ all. guys have ideas on how solve this. the code used. import system import javax.print import java.io java.lang import thread filestream = java.io.bufferedinputstream(java.io.fileinputstream("c:/geeks.txt")) psinformat = javax.print.docflavor.input_stream.autosense mydoc = javax.print.simpledoc(filestream,psinformat,none) aset = javax.print.attribute.hashprintrequestattributeset() aset.add(javax.print.attribute.standard.copies(2)) services = javax.print.printservicelookup.lookupdefaultprintservice() job = services.createprintjob() job.print(mydoc, none) thread.sleep(20000) filestream.close() .docx files have far more formatting , nonsense going on text file. open 1 using notepad , you'll see i'm talking about. because of this, you'll need use dedicated library read docx file, such https://pypi.python.org/pypi/docx/0.2.4 ...
Read more

qml - Is it possible to implement SystemTrayIcon functionality in Qt Quick application -

-
i writing qtquick desktop application. use both c++ (for functionality) , qml (for ui) in it. use qquickview show interface written in qml. i want application reside in system tray when minimised. i mean functionality similar example. http://qt-project.org/doc/qt-4.8/desktop-systray.html . trying implement feature not find way in qt quick application. here main.cpp code: #include <qguiapplication> #include <qqmlengine> #include <qqmlcontext> #include <qqmlfileselector> #include <qquickview> #include "myapp.h" int main(int argc, char* argv[]) { qguiapplication app(argc,argv); app.setapplicationname(qfileinfo(app.applicationfilepath()).basename()); qdir::setcurrent(qapp->applicationdirpath()); myapp myappobject; qquickview view; view.connect(view.engine(), signal(quit()), &app, slot(quit())); view.rootcontext()->setcontextproperty("myappobject", &myappobject); new qqmlfilese...
Read more

c# - How to use process.StartInfo.Arguments in command prompt? -

-
i wanted use c# visual studio open command prompt using process.startinfo.filename , process.startinfo.arguments give commands automatically not work out. knows how this? ( using cmd call python script ) try using following. hope helps. replace notepad.exe else like. process p = new process(); p.startinfo = new processstartinfo("cmd.exe", "/c notepad.exe"); p.start(); please refer http://ss64.com/nt/cmd.html more details.
Read more