Re: Hivemall ffm classifier and class weights

2019-11-01 Thread Shadi Mari
But class weighting doesnt need recalibration, isnt? Shadi On Friday, November 1, 2019, Makoto Yui wrote: > True. And, I think oversampling is better than class weighting for > accuracy. > > Makoto > > 2019年11月1日(金) 19:37 Shadi Mari : > >> Thanks for the explanation

Re: Hivemall ffm classifier and class weights

2019-10-31 Thread Shadi Mari
t; > Sckit SGD [1] adjust y as follows: > "n_samples / (n_classes * np.bincount(y))" > [1] > https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDClassifier.html > > I think this can easily be achieved using SQL. > > Thanks, > Makoto >

Hivemall ffm classifier and class weights

2019-10-31 Thread Shadi Mari
Hello I am having extremely imbalanced dataset and trying to find support for class weights in hivemall ffm classifier in specific, however couldnt find any mention in the docs. Is this feature supported, otherwise i had to go with negative downsampling. Please advise Thank you

Re: Hivemall FFM & Criteo Dataset - LogLess counter

2018-03-04 Thread Shadi Mari
Hi Maktoto, It seems priorities have been shifted a bit since our last discussion. I was wondering if you have thoughts when the FFM prediction logic will stabilize and get mature? I am referring to this issue https://issues.apache.org/jira/browse/HIVEMALL-24 which is planned to be part of 0.5.2

Re: Logistic Regression + FTRL/AdaGrad

2017-10-24 Thread Shadi Mari
userguide/binaryclass/general.html > > Find options by > > select > train_classifier(features, label, '-help') as (feature, weight) > from > a9a_train > > Makoto > > 2017-10-25 0:30 GMT+09:00 Shadi Mari <shadim...@gmail.com>: > > Hello, > > > > I

Logistic Regression + FTRL/AdaGrad

2017-10-24 Thread Shadi Mari
Hello, I am looking to use HiveMALL LR for CTR prediction on Criteo dataset. I can find that LR + SGD is supported by default. Any idea how to use LR along with FTRL or AdaGrad optimizer ? Many Thanks

Re: Hivemall FFM & Criteo Dataset - LogLess counter

2017-10-18 Thread Shadi Mari
> > initial losses. > > > > Further investigation is required but I need to focus on the first > > Apache release for this month. > > > > GA of FFM will be v0.5.1 release scheduled on Dec. > > > > Makoto > > > > 2017-10-18 1:36 GMT+09:00 Makoto

Re: [nhusers] Lookups and NH 2nd Level Caching

2011-06-07 Thread Shadi Mari
anybody knows the answer for such scenario? On Mon, Jun 6, 2011 at 2:27 PM, Shadi Mari shadim...@gmail.com wrote: Hello, I have the following domain definitions: class LookupT { int Id; public string Name { get { return based on culture; }} private string Name_en; private string

[nhusers] Lookups and NH 2nd Level Caching

2011-06-06 Thread Shadi Mari
Hello, I have the following domain definitions: class LookupT { int Id; public string Name { get { return based on culture; }} private string Name_en; private string Name_fr; } class Country : LookupCountry { } class Employee { int Id; float Salray; Country Country; } I also have a db table