Re: [Scikit-learn-general] Small bug/inconsistency in sklearn.feature_extraction.text.CountVectorizer

On Thu, Mar 28, 2013 at 8:37 AM, Lars Buitinck wrote: > There's already a pull request that speeds up CountVectorizer and > returns a csr_matrix. I think we should merge it in soon. Actually, it returns a CSC matrix and we were arguing whether it should be the default output or not. I think that

Re: [Scikit-learn-general] Questions about converting categorical data into input data for an SVM

2013/3/27 Anne Dwyer : > Just to clarify, you are saying that there is no procedure in scikit that > will transform categorical feature values into numerical values like I was > trying to do here. Correct? Not that I know of. DictVectorizer comes quite close, though. -- Lars Buitinck Scientific

Re: [Scikit-learn-general] Small bug/inconsistency in sklearn.feature_extraction.text.CountVectorizer

2013/3/27 Tom Fawcett : > I’ve identified a bug/inconsistency in sklearn.feature_extraction.text. > TfidfVectorizer returns a matrix of type scipy.sparse.csr.csr_matrix; whereas > CountVectorizer returns scipy.sparse.coo.coo_matrix, which don’t support > multiple (array) indexing. > > Below is a

[Scikit-learn-general] Small bug/inconsistency in sklearn.feature_extraction.text.CountVectorizer

I’ve identified a bug/inconsistency in sklearn.feature_extraction.text. TfidfVectorizer returns a matrix of type scipy.sparse.csr.csr_matrix; whereas CountVectorizer returns scipy.sparse.coo.coo_matrix, which don’t support multiple (array) indexing. Below is a short (silly) example that demonstr

Re: [Scikit-learn-general] Questions about converting categorical data into input data for an SVM

Thanks for your answer. Just to clarify, you are saying that there is no procedure in scikit that will transform categorical feature values into numerical values like I was trying to do here. Correct? Anne Dwyer On Wed, Mar 27, 2013 at 4:05 PM, Lars Buitinck wrote: > 2013/3/27 Anne Dwyer : > >

[Scikit-learn-general] Sparse Data Format for SVM SCIKIT

Hello all Is there any example of using datasets.load_svmlight_file() already available or any data set in sparse matrix format to be used along with this function in SCIKIT? Regards -- Own the Future-Intel® Level Up Gam

Re: [Scikit-learn-general] Questions about converting categorical data into input data for an SVM

2013/3/27 Anne Dwyer : > I'm trying to convert categorical data to input data for an SVM. I am trying > to transform the data first to label encoded data, then use the one hot > encoding procedure. > > Can some please explain what I am doing wrong? LabelEncoder is for class labels, not features. I

[Scikit-learn-general] Questions about converting categorical data into input data for an SVM

I'm trying to convert categorical data to input data for an SVM. I am trying to transform the data first to label encoded data, then use the one hot encoding procedure. Here's a subset of the data that I'm trying to transform: array_both= [['3' 'male'] ['1' 'female'] ['3' 'female'] ['1' 'female