This was mainly due to the detection of a numerical feature as a categorical one.
We suggested increasing the categorical threshold as a work-around. @thushan did it work? On Tue, Aug 11, 2015 at 5:50 PM, Thushan Ganegedara <[email protected]> wrote: > This issue occurs, if I turn the response variable to a categorical > variable. If I get the variable as a numerical variable, the values are > read correctly. > > So I presume there is a fault in categorical conversion of the variable. > > On Tue, Aug 11, 2015 at 7:11 PM, Thushan Ganegedara <[email protected]> > wrote: > >> I still get the same result >> >> 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 >> 1.0 1.0 1.0 12.0 12.0 12.0 12.0 12.0 12.0 >> 12.0 12.0 12.0 12.0 13.0 13.0 13.0 13.0 13.0 13.0 >> 13.0 13.0 13.0 13.0 14.0 14.0 14.0 14.0 14.0 >> 14.0 14.0 14.0 15.0 15.0 15.0 15.0 15.0 15.0 >> 15.0 15.0 15.0 15.0 15.0 15.0 16.0 16.0 16.0 16.0 >> 16.0 16.0 16.0 16.0 17.0 17.0 17.0 17.0 17.0 >> 17.0 17.0 17.0 17.0 17.0 18.0 18.0 18.0 18.0 >> 18.0 18.0 18.0 18.0 18.0 18.0 18.0 19.0 19.0 19.0 >> 19.0 19.0 19.0 19.0 19.0 19.0 19.0 19.0 19.0 >> 19.0 19.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 >> 2.0 2.0 2.0 2.0 2.0 2.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 4.0 4.0 4.0 4.0 4.0 4.0 >> 4.0 4.0 4.0 4.0 4.0 4.0 5.0 5.0 5.0 5.0 >> 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 >> 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 >> 6.0 6.0 6.0 7.0 7.0 7.0 7.0 7.0 7.0 7.0 >> 7.0 7.0 7.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >> 3.0 3.0 3.0 3.0 >> >> On Tue, Aug 11, 2015 at 7:05 PM, Nirmal Fernando <[email protected]> wrote: >> >>> Can you use following code and try; >>> >>> List<LabeledPoint> points = labeledPoints.collect(); >>> for(int i=0;i<points.size();i++){ >>> System.out.print(points.get(i).label() + "\t"); >>> } >>> >>> On Tue, Aug 11, 2015 at 2:30 PM, Thushan Ganegedara <[email protected]> >>> wrote: >>> >>>> I used the following snippet >>>> >>>> for(int i=0;i<labeledPoints.collect().size();i++){ >>>> System.out.print(labeledPoints.collect().get(i).label() + >>>> "\t"); >>>> } >>>> >>>> in the public MLModel build() throws MLModelBuilderException in >>>> DeeplearningModelBuilder.java >>>> >>>> >>>> On Tue, Aug 11, 2015 at 6:17 PM, Nirmal Fernando <[email protected]> >>>> wrote: >>>> >>>>> Hi thushan, >>>>> >>>>> We need more info. What did you exactly print and where? >>>>> >>>>> On Tue, Aug 11, 2015 at 12:47 PM, Thushan Ganegedara <[email protected] >>>>> > wrote: >>>>> >>>>>> Hi, >>>>>> >>>>>> I found the potential cause of the poor accuracy for the leaf >>>>>> dataset. It seems the data read into ML is wrong. >>>>>> >>>>>> I have attached the data file as a CSV (classes are in the last >>>>>> column) >>>>>> >>>>>> However, when I print out the labels of the read data (classes), it >>>>>> looks something like below. Clearly there aren't this many "3.0" classes >>>>>> and there should be classes up to 36.0. >>>>>> >>>>>> Is this caused by a bug? >>>>>> >>>>>> 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 >>>>>> 1.0 1.0 1.0 1.0 12.0 12.0 12.0 12.0 12.0 >>>>>> 12.0 12.0 12.0 12.0 12.0 13.0 13.0 13.0 13.0 >>>>>> 13.0 13.0 >>>>>> 13.0 13.0 13.0 13.0 14.0 14.0 14.0 14.0 >>>>>> 14.0 14.0 14.0 14.0 15.0 15.0 15.0 15.0 15.0 >>>>>> 15.0 15.0 15.0 15.0 15.0 15.0 15.0 16.0 16.0 >>>>>> 16.0 16.0 >>>>>> 16.0 16.0 16.0 16.0 17.0 17.0 17.0 17.0 >>>>>> 17.0 17.0 17.0 17.0 17.0 17.0 18.0 18.0 18.0 >>>>>> 18.0 18.0 18.0 18.0 18.0 18.0 18.0 18.0 19.0 >>>>>> 19.0 19.0 >>>>>> 19.0 19.0 19.0 19.0 19.0 19.0 19.0 19.0 >>>>>> 19.0 19.0 19.0 2.0 2.0 2.0 2.0 2.0 2.0 >>>>>> 2.0 2.0 2.0 2.0 2.0 2.0 2.0 3.0 3.0 >>>>>> 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 4.0 4.0 4.0 4.0 4.0 >>>>>> 4.0 4.0 4.0 4.0 4.0 4.0 4.0 5.0 5.0 >>>>>> 5.0 5.0 >>>>>> 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 >>>>>> 5.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 >>>>>> 6.0 6.0 6.0 6.0 7.0 7.0 7.0 7.0 7.0 >>>>>> 7.0 7.0 >>>>>> 7.0 7.0 7.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 >>>>>> 3.0 3.0 >>>>>> 3.0 3.0 3.0 3.0 >>>>>> >>>>>> -- >>>>>> Regards, >>>>>> >>>>>> Thushan Ganegedara >>>>>> School of IT >>>>>> University of Sydney, Australia >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> >>>>> Thanks & regards, >>>>> Nirmal >>>>> >>>>> Team Lead - WSO2 Machine Learner >>>>> Associate Technical Lead - Data Technologies Team, WSO2 Inc. >>>>> Mobile: +94715779733 >>>>> Blog: http://nirmalfdo.blogspot.com/ >>>>> >>>>> >>>>> >>>> >>>> >>>> -- >>>> Regards, >>>> >>>> Thushan Ganegedara >>>> School of IT >>>> University of Sydney, Australia >>>> >>> >>> >>> >>> -- >>> >>> Thanks & regards, >>> Nirmal >>> >>> Team Lead - WSO2 Machine Learner >>> Associate Technical Lead - Data Technologies Team, WSO2 Inc. >>> Mobile: +94715779733 >>> Blog: http://nirmalfdo.blogspot.com/ >>> >>> >>> >> >> >> -- >> Regards, >> >> Thushan Ganegedara >> School of IT >> University of Sydney, Australia >> > > > > -- > Regards, > > Thushan Ganegedara > School of IT > University of Sydney, Australia > -- Thanks & regards, Nirmal Team Lead - WSO2 Machine Learner Associate Technical Lead - Data Technologies Team, WSO2 Inc. Mobile: +94715779733 Blog: http://nirmalfdo.blogspot.com/
_______________________________________________ Dev mailing list [email protected] http://wso2.org/cgi-bin/mailman/listinfo/dev
