Re: [Scikit-learn-general] cannot import MLPClassifier

Hi, Herbert, one way would be to go to the github repo, fork it (or download the zip file), go to its main dir and install it via "python setup.py install” (you probably want to do this is a separate virtual environment so that you can toggle between your different scikit installations) Haven’t

Re: [Scikit-learn-general] cannot import MLPClassifier

Hi Herbert, You'll need to download and compile the source code from github for the dev version. Here's some instructions: - http://scikit-learn.org/stable/install.html - http://scikit-learn.org/stable/developers/contributing.html#retrieving-the-latest-code On Mon, Dec 7, 2015 at 7:03 PM, Herb

[Scikit-learn-general] cannot import MLPClassifier

Hello everyone, I just wanted to use the MLPClassifier from Scikit-learn, but I'm getting an import Error: ImportError: cannot import name MLPClassifier. I read from one post on stack overflow that it is only available on scikit-learn 0.18 dev version. Is there a way to get the 0.18 dev versio

Re: [Scikit-learn-general] Dynamic Time Warping Contribution

I don't have any DTW code written, but I could definitely prototype how a lower bound callable might get incorporated into sklearn On Mon, Dec 7, 2015 at 5:28 PM, Stéfan van der Walt wrote: > On Mon, Dec 7, 2015 at 6:04 AM, Gael Varoquaux > wrote: > >> hello pandora's box ;) > >> I thought we d

Re: [Scikit-learn-general] Dynamic Time Warping Contribution

On Mon, Dec 7, 2015 at 6:04 AM, Gael Varoquaux wrote: >> hello pandora's box ;) >> I thought we don't want to have time-series specific code? > > I agree. We should strive to make something like this pluggeable into > scikit-learn, but not have in inside. If you are seeking a home for DTW code, w

Re: [Scikit-learn-general] Analyzer and tokenizer in (Count/TfIdf)Vectorizer

In the case of "char_wb" it sounds indeed like a custom tokenizer should be called if given. That would require a different implementation than the current one, however. You might want to file an issue. Sebastian's suggestion works, but note that scikit-learn's default tokenization is not the same

Re: [Scikit-learn-general] Analyzer and tokenizer in (Count/TfIdf)Vectorizer

Hi, I think for most implementations you need to tokenize your text into a 1-gram vector prior to stemming/lemmatization, but you can always call " ".join or so if you'd like to get the results back as a string. >>> sentence = 'Stemming is funnier than a bummer says the sushi loving >>> comput

Re: [Scikit-learn-general] Analyzer and tokenizer in (Count/TfIdf)Vectorizer

Hi. I would say what you are doing with lemmatization is not tokenization but preprocessing. You are not creating tokens, right? The tokens are the char n-grams. So what is the problem in using the preprocessing option? I'm not super familiar with the NLP lingo, though, so I might be missing

Re: [Scikit-learn-general] Dynamic Time Warping Contribution

I would say prototype it and let's see what it implies on the code. A -- Go from Idea to Many App Stores Faster with Intel(R) XDK Give your users amazing mobile app experiences with Intel(R) XDK. Use one codebase in this

Re: [Scikit-learn-general] Dynamic Time Warping Contribution

What about adding the option for users to pass in a callable "lower bound" function to a nearest neighbor search? Then users could use things like the LB Keogh lower bound On Mon, Dec 7, 2015 at 9:04 AM, Gael Varoquaux < gael.varoqu...@normalesup.org> wrote: > > hello pandora's box ;) > > I thoug

Re: [Scikit-learn-general] Dynamic Time Warping Contribution

> hello pandora's box ;) > I thought we don't want to have time-series specific code? I agree. We should strive to make something like this pluggeable into scikit-learn, but not have in inside. -- Go from Idea to Many App

Re: [Scikit-learn-general] Dynamic Time Warping Contribution

On 12/07/2015 04:33 AM, Alexandre Gramfort wrote: >> How do you plan to represent variable-length time series? Lists of 1d numpy >> arrays work but would be slow I guess. The ideal representation needs to be >> compatible with grid search and fast. > good point. I was thinking of forcing all time s

[Scikit-learn-general] Regarding Research in Big Data

Hi, I'm doing research in big data mining. I'm new to python. Can anyone suggest tool for big data analytics/mining in python? -- - Regards, PROF. CHINTAN BHATT PhD*, Computer Science https://sites.google.com/a/ecchanga.ac.in/chintan/ ---

Re: [Scikit-learn-general] Dynamic Time Warping Contribution

> How do you plan to represent variable-length time series? Lists of 1d numpy > arrays work but would be slow I guess. The ideal representation needs to be > compatible with grid search and fast. good point. I was thinking of forcing all time series to have the same length. or use dtype = object l

Re: [Scikit-learn-general] Dynamic Time Warping Contribution

How do you plan to represent variable-length time series? Lists of 1d numpy arrays work but would be slow I guess. The ideal representation needs to be compatible with grid search and fast. Mathieu On Mon, Dec 7, 2015 at 10:35 AM, Dan Shiebler wrote: > Hello, > > I’m not sure if this is the cor

Re: [Scikit-learn-general] Dynamic Time Warping Contribution

> In addition, users > cannot take advantage of the LB Keogh lower bound of dynamic time warping, > which can dramatically speed up the nearest neighbors search. can you give more details on how you would use this in our NN code? if passing a dtw callable is not good enough, that can justify an a