Google Prediction API always returns a score of 1.0 or 0.0 -

-

i'm using the insert function of google-api-ruby-client , passing in array of traininginstances.

when call predict used scores 0.653264 , 0.346736 (using 2 output categories).

after doing refactoring scores of 1.0 , 0.0 , no probabilities in between.

the prediction accuracy (measured me) still close accuracy given status function looks trained model working. want know why i'm not getting scores in between 1 , 0 use in application.

any insight appreciated.

here sample of training data i'm using:

"completed registration",32.54,2,0,1,12 "completed registration",27.05,2,0,1,7 "completed registration",27.29,1,0,3,7 "completed registration",24.15,1,0,1,6 "completed registration",26.36,2,0,1,6 "completed registration",27.0,3,1,3,5 "completed registration",22.15,3,1,3,5 "completed registration",27.9,2,1,1,1 "completed registration",21.21,2,0,3,0 "completed registration",24.36,3,0,3,0 "completed registration",53.52,2,1,3,0 "completed registration",24.15,8,0,3,-6 "cancelled registration",,2,0,3,76 "cancelled registration",,2,0,3,76 "cancelled registration",,3,1,3,69 "cancelled registration",38.27,5,0,1,63 "cancelled registration",37.25,2,0,3,50 "cancelled registration",35.59,5,1,3,49 "cancelled registration",19.52,2,0,3,30 "cancelled registration",26.7,2,0,3,28

i've run tests on google api explorer , uploaded training data in csv file.

the data first uploaded (a few weeks ago) results in range of scores. exact same data uploaded today results in scores of 1 or 0.

could character encoding problem?

in case other people have same question i'm posting answer here. turns out internally api can use different classifiers depending on training data. looks difference caused change api , not refactoring.

here explanation @ google https://groups.google.com/forum/#!topic/prediction-api-discuss/7qphonkhzjc


Comments

Post a Comment

Popular posts from this blog

python - Specify path of savefig with pylab or matplotlib -

-
i struggling on how find correct way specify save path (or repository) when calling function savefig in matplotlib or pylab. i tried several syntaxes, everytime python console returns : filenotfounderror: [errno 2] no such file or directory: '../mydocs/resource/frames/myimage.png' currently, have written following : pylab.savefig('../mydocs/resource/frames/myimage.png') does know how it? thanks in advance! the tilde operator , variable $home entered strings , therefore not function while saving. have provide either relative path (as did) or provide full path. eg. pylab.savefig("/home/username/desktop/myfig.png") .
Read more

How to run C# code using mono without Xamarin in Android? -

-
i want create program uses mono runtime run c# code (binary not application). i copied libmono.so unity app , i'm write code using documentation: http://www.mono-project.com/docs/advanced/embedding/ and it's working have problems references can access classes mscorlib.dll , have problems system.reflection.assembly class throwing exception, when print console see: system.missingmethodexception: method not found: 'system.reflection.assembly.op_inequality' can fix it? can use mono runtime in c binary on android? ps. i'm using android tests. edit: found solution loading libraries, need set environment variable mono_path libraries path. but have still problems assembly class. the system.missingmethodexception error have seems .net profile mismatch error, 1 explained here . can use mono runtime in c binary on android? yes, using ndk guess. there alternatives can use instead of rolling own, such xmadevlab .
Read more

c# - SharpSsh Command Execution -

-
Image
i trying create ssh client using c#. using tamir.sharpssh library. having issues sending command , getting appropriate response server. the way have been testing sending on ls command, followed cd.. command, followed ls command again -- see whether or not cd.. command being executed properly. however, there no difference between first ls command , second ls command. not entirely sure doing wrong. perhaps using wrong ssh type. have provided code , output getting. have added output expect. using tamir.sharpssh; namespace sshnetexample { class program { static void main(string[] args) { sshexec ssh = null; try { console.write("-connecting..."); ssh = new sshexec(host, username, password); ssh.connect(); console.writeline("ok ({0}/{1})", ssh.cipher, ssh.mac); console.writeline("server version={0}, client versio...
Read more