{ "id": 6803, "title": "Beating the Benchmark", "url": "/competitions/galaxy-zoo-the-galaxy-challenge/discussion/6803", "author_name": "", "post_date": "2014-01-10T15:45:26.303Z", "votes": 15, "comment_count": 29, "views": 12583, "content": "

Since, I've been ridiculed previously for the beating the benchmark posts, this time I wont post the code, but a method which would enable you to beat the central pixel benchmark and would take only 20mins.

\n

 

\n

Step 1 : for all images in train and test, resize to 50x50, and vectorize to 1-D array.

\n

Step 2: Run the RandomForestRegression from scikit-learn on the train and test set with 10 estimators.

\n

Step 3: Submit Results

\n

Step 4: You have beaten the benchmark

\n

Step 5: Click thanks if this post helped you ;)

", "messages": [ { "id": "37233", "postDate": "01/10/2014 15:45:26", "content": "

Since, I've been ridiculed previously for the beating the benchmark posts, this time I wont post the code, but a method which would enable you to beat the central pixel benchmark and would take only 20mins.

\n

 

\n

Step 1 : for all images in train and test, resize to 50x50, and vectorize to 1-D array.

\n

Step 2: Run the RandomForestRegression from scikit-learn on the train and test set with 10 estimators.

\n

Step 3: Submit Results

\n

Step 4: You have beaten the benchmark

\n

Step 5: Click thanks if this post helped you ;)

", "rawMarkdown": "", "votes": null }, { "id": "37245", "postDate": "01/10/2014 18:35:54", "content": "

[quote=Abhishek;37233]

\n

Since, I've been ridiculed previously for the beating the benchmark posts, 

\n

[/quote]

\n

LOL.

", "rawMarkdown": "", "votes": null }, { "id": "37258", "postDate": "01/10/2014 21:28:08", "content": "

aint that true :P 

\n

[quote=Giulio;37245]

\n

[quote=Abhishek;37233]

\n

Since, I've been ridiculed previously for the beating the benchmark posts, 

\n

[/quote]

\n

LOL.

\n

[/quote]

", "rawMarkdown": "", "votes": null }, { "id": "37259", "postDate": "01/10/2014 21:35:20", "content": "

[quote=Abhishek;37258]

\n

aint that true :P 

\n

[quote=Giulio;37245]

\n

[quote=Abhishek;37233]

\n

Since, I've been ridiculed previously for the beating the benchmark posts, 

\n

[/quote]

\n

LOL.

\n

[/quote]

\n

[/quote]

\n

C'mon! You ended up with more fans than haters :-)

", "rawMarkdown": "", "votes": null }, { "id": "37269", "postDate": "01/11/2014 04:28:06", "content": "

I've never done image processing before. Would it be possible to post a sample code on how to resize and vectorize it to 1-D array? 

\n

 

\n

Thanks.

", "rawMarkdown": "", "votes": null }, { "id": "37275", "postDate": "01/11/2014 09:36:20", "content": "

import numpy as np
import pandas as pdimport cv2

\n

images = []
image_files = sorted(os.listdir('images_training'))
image_files = ["{}/{}".format(name, f) for f in image_files]for imgf in image_files:
    img = cv2.imread(imgf, 0)
    img = cv2.resize(img, (128, 128), interpolation=cv2.INTER_CUBIC)
    length = np.prod(img.shape)
    img = np.reshape(img, length)
    images.append(img)

\n

images = np.vstack(images)

\n

Now images is feature matrix which you can use with scikit learn algorithms.

", "rawMarkdown": "", "votes": null }, { "id": "37320", "postDate": "01/12/2014 17:57:18", "content": "

Thanks Michal, but I can't find the cv2 module. I tried to install via conda and pip, but it didn't work. 

", "rawMarkdown": "", "votes": null }, { "id": "37321", "postDate": "01/12/2014 18:48:11", "content": "

What OS do you use? If it is Mac - there seems to be lot of challenges in making it work. Linux has been the most straight forward. 

", "rawMarkdown": "", "votes": null }, { "id": "37322", "postDate": "01/12/2014 18:49:03", "content": "

Meanwhile - a good resource for processing images using Python

\n

http://programmingcomputervision.com/

", "rawMarkdown": "", "votes": null }, { "id": "37323", "postDate": "01/12/2014 19:17:05", "content": "

[quote=Frank Schilder;37320]

\n

Thanks Michal, but I can't find the cv2 module. I tried to install via conda and pip, but it didn't work. 

\n

[/quote]

\n

cv2 refers to python module as provided by OpenCV. (http://opencv.org/)

\n

Installations steps for Windows are documented here:

\n

http://docs.opencv.org/trunk/doc/py_tutorials/py_setup/py_setup_in_windows/py_setup_in_windows.html

", "rawMarkdown": "", "votes": null }, { "id": "37326", "postDate": "01/13/2014 02:26:48", "content": "

If you're Mac, your probably best off with the macports build of OpenCV.  Installation of the dependencies can be a bit frustrating - there's a good tutorial here:

\n

http://compphotography.wordpress.com/2013/04/08/macos-python-install-macports/

\n

 

\n

Also, to go along with the code example, if you want to vectorize a single image as an array, you can use the ravel() method - which will combine all

\n

vector = cv2.imread('image.jpg').ravel()

", "rawMarkdown": "", "votes": null }, { "id": "37329", "postDate": "01/13/2014 05:09:17", "content": "

I was able to download and install openCV, but my machine is a mac and installing the cv2 module seemed rather painful judging by the web pages offering various solutions to this problem. I have been using PIL so far, but it takes quite a while to load the images and I was hoping cv2 would be faster. 

\n

The question is whether it is worthwhile to invest the extra time to install cv2. OpenCV seems to offer quite a lot of useful resources for image processing though. Anybody out there who was able to run cv2/OpenCV on a Mac?

\n

[I hadn't seen the post by Raymond Klass when I wrote this, I'll try macports then, although I have been using brew and pip lately.]

", "rawMarkdown": "", "votes": null }, { "id": "37330", "postDate": "01/13/2014 05:42:26", "content": "

[quote=Frank Schilder;37329]

\n

The question is whether it is worthwhile to invest the extra time to install cv2. OpenCV seems to offer quite a lot of useful resources for image processing though. Anybody out there who was able to run cv2/OpenCV on a Mac?

\n

[/quote]

\n

 

\n

I work on a mac and installed opencv using homebrew. 

", "rawMarkdown": "", "votes": null }, { "id": "37337", "postDate": "01/13/2014 09:55:53", "content": "

Abhishek - what MacOS are you using? Out of interest, have you manged to get Theano and Pylearn2 working with GPU? I had to install Anaconda Python to get Python on my Mac. I'm on Mac OS X Lion 10.7.5 (11G63) - there was a conflict with scipy (xcodes/gcc/lvcc ?  etc) when I tried to install packages individually

", "rawMarkdown": "", "votes": null }, { "id": "37339", "postDate": "01/13/2014 10:11:17", "content": "

Im using OSX mavericks. I didnt face any conflicts with scipy and all. As I dont have a GPU on my mac, I cannot use theano/pylearn2 for convolutional neural networks. However, I can use neural networks which do not require GPU and thus, theano/pylearn2 are working fine for me. I had a hard time installing opencv, but I finally managed to install it. (https://github.com/Homebrew/homebrew-science/issues/402)

\n

[quote=Domcastro;37337]

\n

Abhishek - what MacOS are you using? Out of interest, have you manged to get Theano and Pylearn2 working with GPU? I had to install Anaconda Python to get Python on my Mac. I'm on Mac OS X Lion 10.7.5 (11G63) - there was a conflict with scipy (xcodes/gcc/lvcc ?  etc)

\n

[/quote]

", "rawMarkdown": "", "votes": null }, { "id": "37340", "postDate": "01/13/2014 10:20:30", "content": "

I use Anaconda on Mac. And this did the trick for opencv. A bit of effort though ! 

\n

https://gist.github.com/welch/6468594

", "rawMarkdown": "", "votes": null }, { "id": "37342", "postDate": "01/13/2014 11:15:38", "content": "

@domcastro, did you install an external GPU for your Mac? 

\n

[quote=Domcastro;37337]

\n

Abhishek - what MacOS are you using? Out of interest, have you manged to get Theano and Pylearn2 working with GPU? I had to install Anaconda Python to get Python on my Mac. I'm on Mac OS X Lion 10.7.5 (11G63) - there was a conflict with scipy (xcodes/gcc/lvcc ?  etc) when I tried to install packages individually

\n

[/quote]

", "rawMarkdown": "", "votes": null }, { "id": "37347", "postDate": "01/13/2014 13:32:42", "content": "

ah no- I have set up Theano and Pylearn2 on the Mac but uses CPU.  I have a 64 bit windows machine too so bought a GEforce graphics card and installed that. Having trouble getting theano to work on winows though

", "rawMarkdown": "", "votes": null }, { "id": "37483", "postDate": "01/17/2014 00:01:04", "content": "

Check out http://scikit-image.org/ . It has got easy to use functional API in Python that work with bare Numpy arrays. To start, check out their Examples section : http://scikit-image.org/docs/dev/auto_examples/

\n

Github : https://github.com/scikit-image/scikit-image

", "rawMarkdown": "", "votes": null }, { "id": "37524", "postDate": "01/17/2014 19:07:29", "content": "

Hi there, any sample code using R?

\n

Thanks a lot,

", "rawMarkdown": "", "votes": null }, { "id": "37648", "postDate": "01/20/2014 20:24:00", "content": "

Did you use the values directly from the 'training_solutions.csv' file as the target label vector in sklearn? 

", "rawMarkdown": "", "votes": null }, { "id": "37925", "postDate": "01/25/2014 17:11:55", "content": "

Very helpful, thanks. I'm familiar with the idea behind random forests, but it seems like training them is very slow in scikit-learn. Is that normal? I'm using 56x56 greyscale samples. I've been setting max_depth to around 5, max_features to sqrt, min_samples_split to 20 and that helps. But if I want to have say 80 trees it can take a minute or two.

", "rawMarkdown": "", "votes": null }, { "id": "37930", "postDate": "01/25/2014 23:47:28", "content": "

I also had a lot of trouble getting opencv and related tools to work on my Mac. I was finally able to do it quiet painlessly by using Macports. 

\n

If you suspect you have a foobared version of macports installed, or an install thats really old and hasn't been updated in a while you might think about removing it using the procedure in the following page http://guide.macports.org/chunked/installing.macports.uninstalling.html.

\n

Install the latest version of Macports and then run the following on the command line (I found it also helps if you are running Mavericks):

\n

sudo port install python27

\n

sudo port select --set python python27

\n

sudo port install qt4

\n

sudo port install py27-matplotlib py27-pil py27-scipy py27-pyside py27-numpy

\n

#To get matplotlib to work properly you will want to change its configuration to use QT.

\n

# Do this by going to (your home dir)/.matplotlib/matplotlibrc and change the backend parameter to QT4Agg. Then change the backend.qt4 parameter to PySide. You might need to run python and import matplotlib to get the .matplotlib directory to pop up.

\n

sudo port install opencv +python27

\n

#At this point test your opencv install by opening the python prompt and import cv2.

\n

--- Additional helpful tools ---

\n

sudo port install py27-setuptools

\n

#you will want to go to the /opt/local/bin (where macports stores all the executables) directory and ln -s easy_install-2.7 easy_install

\n

sudo port install py27-ipython

\n

#look into using python notebook ;) you will need to install additional packages to get this to work.

", "rawMarkdown": "", "votes": null }, { "id": "38459", "postDate": "02/06/2014 22:26:58", "content": "

Hello!

\n

Thanks for the tip of using RandomForestRegressor!

\n

I'm also using Mac OS X Mavericks and Scikit-learn.

\n

But I think I have some optimization problems... how much time do you spend fitting the training data??

\n

 

\n

Regards!

", "rawMarkdown": "", "votes": null }, { "id": "38479", "postDate": "02/07/2014 15:29:30", "content": "

Building a model over here takes two or three days. I guess we are all on the same boat.

", "rawMarkdown": "", "votes": null }, { "id": "38618", "postDate": "02/11/2014 07:45:44", "content": "

How can you use a 1D vector to predict the output which is also a vector?

\n

We have these images x in X, and we have the target output which is a vector y in Y.  The target vectors in Y look like <prediction for class 1.1, prediction for class 1.2, ...>, and you've converted each image x in X to be 1D vectors as well.

\n

You say you used Random Forest Regression to make a predictive model from X to Y.  But with Random Forest, you need the Y to be either a real number for regression or a class for classification -- it can't be a vector.

\n

So when you build your model, are you actually building a new forest to classify every entry of the y vector in other words, growing a new forest for every class and subclass?  Or are you focussing your efforts on a classification of just class 1 and zeroing everything else?

\n

Thanks a lot, Abhishek.

", "rawMarkdown": "", "votes": null }, { "id": "38666", "postDate": "02/12/2014 04:25:43", "content": "

[quote=Richard Craib;38618]

\n

So when you build your model, are you actually building a new forest to classify every entry of the y vector in other words, growing a new forest for every class and subclass?  Or are you focussing your efforts on a classification of just class 1 and zeroing everything else?

\n

[/quote]

\n

It's training a different random forest for each output.  If you're using scikit-learn then most (all?) of the regression and classification algorithms will handle it automatically.  They check the shape of the ndarray and behave differently for one output (1-dim) vs multiple outputs (2-dim).

\n

 

", "rawMarkdown": "", "votes": null }, { "id": "38704", "postDate": "02/13/2014 00:29:35", "content": "

[quote=Keith Trnka;38666]

\n

[quote=Richard Craib;38618]

\n

So when you build your model, are you actually building a new forest to classify every entry of the y vector in other words, growing a new forest for every class and subclass?  Or are you focussing your efforts on a classification of just class 1 and zeroing everything else?

\n

[/quote]

\n

It's training a different random forest for each output.  If you're using scikit-learn then most (all?) of the regression and classification algorithms will handle it automatically.  They check the shape of the ndarray and behave differently for one output (1-dim) vs multiple outputs (2-dim).

\n

 

\n

[/quote]

\n

I have only found ensemble methods like RandomForestRegressor and ExtraTreesRegressor are able to automagically give multiple regression outputs. Are there other options in scikit-learn? I suspect if you want to use other learning algorithms you will need to stitch them together manually.

", "rawMarkdown": "", "votes": null }, { "id": "38714", "postDate": "02/13/2014 05:01:06", "content": "

[quote=Jeremy;38704]

\n

I have only found ensemble methods like RandomForestRegressor and ExtraTreesRegressor are able to automagically give multiple regression outputs. Are there other options in scikit-learn? I suspect if you want to use other learning algorithms you will need to stitch them together manually.

\n

[/quote]

\n

I'm pretty new to scikit-learn, but linear_model.Ridge will automatically work for multiple regression outputs.  The other classes in the linear package probably do too.

\n

I've tried logistic regression but it doesn't support multiple binary classifications automatically.  It looks like sklearn.multiclass.OneVsRestClassifier might be the an option but if you're already dealing with a set of binary classifications you'd need to convert to a single multiple-class output first.

\n

You can check classes individually by reading the doc on the fit function, but I wish there were a high-level table that showed it.

", "rawMarkdown": "", "votes": null }, { "id": "38868", "postDate": "02/15/2014 19:55:25", "content": "

Thanks for the tip. I was able to beat the benchmark using only least squares and a single pixel image matrix, which made the computation time much faster.

\n

 

", "rawMarkdown": "", "votes": null } ], "comments": [ { "id": 37245, "author_name": "adjgiulio", "author_url": "", "post_date": "01/10/2014 18:35:54", "content": "

[quote=Abhishek;37233]

\n

Since, I've been ridiculed previously for the beating the benchmark posts, 

\n

[/quote]

\n

LOL.

", "votes": null, "replies": [] }, { "id": 37258, "author_name": "abhishek", "author_url": "", "post_date": "01/10/2014 21:28:08", "content": "

aint that true :P 

\n

[quote=Giulio;37245]

\n

[quote=Abhishek;37233]

\n

Since, I've been ridiculed previously for the beating the benchmark posts, 

\n

[/quote]

\n

LOL.

\n

[/quote]

", "votes": null, "replies": [] }, { "id": 37259, "author_name": "adjgiulio", "author_url": "", "post_date": "01/10/2014 21:35:20", "content": "

[quote=Abhishek;37258]

\n

aint that true :P 

\n

[quote=Giulio;37245]

\n

[quote=Abhishek;37233]

\n

Since, I've been ridiculed previously for the beating the benchmark posts, 

\n

[/quote]

\n

LOL.

\n

[/quote]

\n

[/quote]

\n

C'mon! You ended up with more fans than haters :-)

", "votes": null, "replies": [] }, { "id": 37269, "author_name": "rouseguy", "author_url": "", "post_date": "01/11/2014 04:28:06", "content": "

I've never done image processing before. Would it be possible to post a sample code on how to resize and vectorize it to 1-D array? 

\n

 

\n

Thanks.

", "votes": null, "replies": [] }, { "id": 37275, "author_name": "mrnustik", "author_url": "", "post_date": "01/11/2014 09:36:20", "content": "

import numpy as np
import pandas as pdimport cv2

\n

images = []
image_files = sorted(os.listdir('images_training'))
image_files = ["{}/{}".format(name, f) for f in image_files]for imgf in image_files:
    img = cv2.imread(imgf, 0)
    img = cv2.resize(img, (128, 128), interpolation=cv2.INTER_CUBIC)
    length = np.prod(img.shape)
    img = np.reshape(img, length)
    images.append(img)

\n

images = np.vstack(images)

\n

Now images is feature matrix which you can use with scikit learn algorithms.

", "votes": null, "replies": [] }, { "id": 37320, "author_name": "fschilder", "author_url": "", "post_date": "01/12/2014 17:57:18", "content": "

Thanks Michal, but I can't find the cv2 module. I tried to install via conda and pip, but it didn't work. 

", "votes": null, "replies": [] }, { "id": 37321, "author_name": "rouseguy", "author_url": "", "post_date": "01/12/2014 18:48:11", "content": "

What OS do you use? If it is Mac - there seems to be lot of challenges in making it work. Linux has been the most straight forward. 

", "votes": null, "replies": [] }, { "id": 37322, "author_name": "rouseguy", "author_url": "", "post_date": "01/12/2014 18:49:03", "content": "

Meanwhile - a good resource for processing images using Python

\n

http://programmingcomputervision.com/

", "votes": null, "replies": [] }, { "id": 37323, "author_name": "mattvonrohr", "author_url": "", "post_date": "01/12/2014 19:17:05", "content": "

[quote=Frank Schilder;37320]

\n

Thanks Michal, but I can't find the cv2 module. I tried to install via conda and pip, but it didn't work. 

\n

[/quote]

\n

cv2 refers to python module as provided by OpenCV. (http://opencv.org/)

\n

Installations steps for Windows are documented here:

\n

http://docs.opencv.org/trunk/doc/py_tutorials/py_setup/py_setup_in_windows/py_setup_in_windows.html

", "votes": null, "replies": [] }, { "id": 37326, "author_name": "raymondklass", "author_url": "", "post_date": "01/13/2014 02:26:48", "content": "

If you're Mac, your probably best off with the macports build of OpenCV.  Installation of the dependencies can be a bit frustrating - there's a good tutorial here:

\n

http://compphotography.wordpress.com/2013/04/08/macos-python-install-macports/

\n

 

\n

Also, to go along with the code example, if you want to vectorize a single image as an array, you can use the ravel() method - which will combine all

\n

vector = cv2.imread('image.jpg').ravel()

", "votes": null, "replies": [] }, { "id": 37329, "author_name": "fschilder", "author_url": "", "post_date": "01/13/2014 05:09:17", "content": "

I was able to download and install openCV, but my machine is a mac and installing the cv2 module seemed rather painful judging by the web pages offering various solutions to this problem. I have been using PIL so far, but it takes quite a while to load the images and I was hoping cv2 would be faster. 

\n

The question is whether it is worthwhile to invest the extra time to install cv2. OpenCV seems to offer quite a lot of useful resources for image processing though. Anybody out there who was able to run cv2/OpenCV on a Mac?

\n

[I hadn't seen the post by Raymond Klass when I wrote this, I'll try macports then, although I have been using brew and pip lately.]

", "votes": null, "replies": [] }, { "id": 37330, "author_name": "abhishek", "author_url": "", "post_date": "01/13/2014 05:42:26", "content": "

[quote=Frank Schilder;37329]

\n

The question is whether it is worthwhile to invest the extra time to install cv2. OpenCV seems to offer quite a lot of useful resources for image processing though. Anybody out there who was able to run cv2/OpenCV on a Mac?

\n

[/quote]

\n

 

\n

I work on a mac and installed opencv using homebrew. 

", "votes": null, "replies": [] }, { "id": 37337, "author_name": "domcastro", "author_url": "", "post_date": "01/13/2014 09:55:53", "content": "

Abhishek - what MacOS are you using? Out of interest, have you manged to get Theano and Pylearn2 working with GPU? I had to install Anaconda Python to get Python on my Mac. I'm on Mac OS X Lion 10.7.5 (11G63) - there was a conflict with scipy (xcodes/gcc/lvcc ?  etc) when I tried to install packages individually

", "votes": null, "replies": [] }, { "id": 37339, "author_name": "abhishek", "author_url": "", "post_date": "01/13/2014 10:11:17", "content": "

Im using OSX mavericks. I didnt face any conflicts with scipy and all. As I dont have a GPU on my mac, I cannot use theano/pylearn2 for convolutional neural networks. However, I can use neural networks which do not require GPU and thus, theano/pylearn2 are working fine for me. I had a hard time installing opencv, but I finally managed to install it. (https://github.com/Homebrew/homebrew-science/issues/402)

\n

[quote=Domcastro;37337]

\n

Abhishek - what MacOS are you using? Out of interest, have you manged to get Theano and Pylearn2 working with GPU? I had to install Anaconda Python to get Python on my Mac. I'm on Mac OS X Lion 10.7.5 (11G63) - there was a conflict with scipy (xcodes/gcc/lvcc ?  etc)

\n

[/quote]

", "votes": null, "replies": [] }, { "id": 37340, "author_name": "rouseguy", "author_url": "", "post_date": "01/13/2014 10:20:30", "content": "

I use Anaconda on Mac. And this did the trick for opencv. A bit of effort though ! 

\n

https://gist.github.com/welch/6468594

", "votes": null, "replies": [] }, { "id": 37342, "author_name": "abhishek", "author_url": "", "post_date": "01/13/2014 11:15:38", "content": "

@domcastro, did you install an external GPU for your Mac? 

\n

[quote=Domcastro;37337]

\n

Abhishek - what MacOS are you using? Out of interest, have you manged to get Theano and Pylearn2 working with GPU? I had to install Anaconda Python to get Python on my Mac. I'm on Mac OS X Lion 10.7.5 (11G63) - there was a conflict with scipy (xcodes/gcc/lvcc ?  etc) when I tried to install packages individually

\n

[/quote]

", "votes": null, "replies": [] }, { "id": 37347, "author_name": "domcastro", "author_url": "", "post_date": "01/13/2014 13:32:42", "content": "

ah no- I have set up Theano and Pylearn2 on the Mac but uses CPU.  I have a 64 bit windows machine too so bought a GEforce graphics card and installed that. Having trouble getting theano to work on winows though

", "votes": null, "replies": [] }, { "id": 37483, "author_name": "ankitagrawal2", "author_url": "", "post_date": "01/17/2014 00:01:04", "content": "

Check out http://scikit-image.org/ . It has got easy to use functional API in Python that work with bare Numpy arrays. To start, check out their Examples section : http://scikit-image.org/docs/dev/auto_examples/

\n

Github : https://github.com/scikit-image/scikit-image

", "votes": null, "replies": [] }, { "id": 37524, "author_name": "vcomas", "author_url": "", "post_date": "01/17/2014 19:07:29", "content": "

Hi there, any sample code using R?

\n

Thanks a lot,

", "votes": null, "replies": [] }, { "id": 37648, "author_name": "gallamine", "author_url": "", "post_date": "01/20/2014 20:24:00", "content": "

Did you use the values directly from the 'training_solutions.csv' file as the target label vector in sklearn? 

", "votes": null, "replies": [] }, { "id": 37925, "author_name": "keithtrnka", "author_url": "", "post_date": "01/25/2014 17:11:55", "content": "

Very helpful, thanks. I'm familiar with the idea behind random forests, but it seems like training them is very slow in scikit-learn. Is that normal? I'm using 56x56 greyscale samples. I've been setting max_depth to around 5, max_features to sqrt, min_samples_split to 20 and that helps. But if I want to have say 80 trees it can take a minute or two.

", "votes": null, "replies": [] }, { "id": 37930, "author_name": "ronlaren", "author_url": "", "post_date": "01/25/2014 23:47:28", "content": "

I also had a lot of trouble getting opencv and related tools to work on my Mac. I was finally able to do it quiet painlessly by using Macports. 

\n

If you suspect you have a foobared version of macports installed, or an install thats really old and hasn't been updated in a while you might think about removing it using the procedure in the following page http://guide.macports.org/chunked/installing.macports.uninstalling.html.

\n

Install the latest version of Macports and then run the following on the command line (I found it also helps if you are running Mavericks):

\n

sudo port install python27

\n

sudo port select --set python python27

\n

sudo port install qt4

\n

sudo port install py27-matplotlib py27-pil py27-scipy py27-pyside py27-numpy

\n

#To get matplotlib to work properly you will want to change its configuration to use QT.

\n

# Do this by going to (your home dir)/.matplotlib/matplotlibrc and change the backend parameter to QT4Agg. Then change the backend.qt4 parameter to PySide. You might need to run python and import matplotlib to get the .matplotlib directory to pop up.

\n

sudo port install opencv +python27

\n

#At this point test your opencv install by opening the python prompt and import cv2.

\n

--- Additional helpful tools ---

\n

sudo port install py27-setuptools

\n

#you will want to go to the /opt/local/bin (where macports stores all the executables) directory and ln -s easy_install-2.7 easy_install

\n

sudo port install py27-ipython

\n

#look into using python notebook ;) you will need to install additional packages to get this to work.

", "votes": null, "replies": [] }, { "id": 38459, "author_name": "leopiney", "author_url": "", "post_date": "02/06/2014 22:26:58", "content": "

Hello!

\n

Thanks for the tip of using RandomForestRegressor!

\n

I'm also using Mac OS X Mavericks and Scikit-learn.

\n

But I think I have some optimization problems... how much time do you spend fitting the training data??

\n

 

\n

Regards!

", "votes": null, "replies": [] }, { "id": 38479, "author_name": "delcacho", "author_url": "", "post_date": "02/07/2014 15:29:30", "content": "

Building a model over here takes two or three days. I guess we are all on the same boat.

", "votes": null, "replies": [] }, { "id": 38618, "author_name": "richardcraib", "author_url": "", "post_date": "02/11/2014 07:45:44", "content": "

How can you use a 1D vector to predict the output which is also a vector?

\n

We have these images x in X, and we have the target output which is a vector y in Y.  The target vectors in Y look like <prediction for class 1.1, prediction for class 1.2, ...>, and you've converted each image x in X to be 1D vectors as well.

\n

You say you used Random Forest Regression to make a predictive model from X to Y.  But with Random Forest, you need the Y to be either a real number for regression or a class for classification -- it can't be a vector.

\n

So when you build your model, are you actually building a new forest to classify every entry of the y vector in other words, growing a new forest for every class and subclass?  Or are you focussing your efforts on a classification of just class 1 and zeroing everything else?

\n

Thanks a lot, Abhishek.

", "votes": null, "replies": [] }, { "id": 38666, "author_name": "keithtrnka", "author_url": "", "post_date": "02/12/2014 04:25:43", "content": "

[quote=Richard Craib;38618]

\n

So when you build your model, are you actually building a new forest to classify every entry of the y vector in other words, growing a new forest for every class and subclass?  Or are you focussing your efforts on a classification of just class 1 and zeroing everything else?

\n

[/quote]

\n

It's training a different random forest for each output.  If you're using scikit-learn then most (all?) of the regression and classification algorithms will handle it automatically.  They check the shape of the ndarray and behave differently for one output (1-dim) vs multiple outputs (2-dim).

\n

 

", "votes": null, "replies": [] }, { "id": 38704, "author_name": "jeremyk", "author_url": "", "post_date": "02/13/2014 00:29:35", "content": "

[quote=Keith Trnka;38666]

\n

[quote=Richard Craib;38618]

\n

So when you build your model, are you actually building a new forest to classify every entry of the y vector in other words, growing a new forest for every class and subclass?  Or are you focussing your efforts on a classification of just class 1 and zeroing everything else?

\n

[/quote]

\n

It's training a different random forest for each output.  If you're using scikit-learn then most (all?) of the regression and classification algorithms will handle it automatically.  They check the shape of the ndarray and behave differently for one output (1-dim) vs multiple outputs (2-dim).

\n

 

\n

[/quote]

\n

I have only found ensemble methods like RandomForestRegressor and ExtraTreesRegressor are able to automagically give multiple regression outputs. Are there other options in scikit-learn? I suspect if you want to use other learning algorithms you will need to stitch them together manually.

", "votes": null, "replies": [] }, { "id": 38714, "author_name": "keithtrnka", "author_url": "", "post_date": "02/13/2014 05:01:06", "content": "

[quote=Jeremy;38704]

\n

I have only found ensemble methods like RandomForestRegressor and ExtraTreesRegressor are able to automagically give multiple regression outputs. Are there other options in scikit-learn? I suspect if you want to use other learning algorithms you will need to stitch them together manually.

\n

[/quote]

\n

I'm pretty new to scikit-learn, but linear_model.Ridge will automatically work for multiple regression outputs.  The other classes in the linear package probably do too.

\n

I've tried logistic regression but it doesn't support multiple binary classifications automatically.  It looks like sklearn.multiclass.OneVsRestClassifier might be the an option but if you're already dealing with a set of binary classifications you'd need to convert to a single multiple-class output first.

\n

You can check classes individually by reading the doc on the fit function, but I wish there were a high-level table that showed it.

", "votes": null, "replies": [] }, { "id": 38868, "author_name": "mattdelhey", "author_url": "", "post_date": "02/15/2014 19:55:25", "content": "

Thanks for the tip. I was able to beat the benchmark using only least squares and a single pixel image matrix, which made the computation time much faster.

\n

 

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "37233": "", "37245": "", "37258": "", "37259": "", "37269": "", "37275": "", "37320": "", "37321": "", "37322": "", "37323": "", "37326": "", "37329": "", "37330": "", "37337": "", "37339": "", "37340": "", "37342": "", "37347": "", "37483": "", "37524": "", "37648": "", "37925": "", "37930": "", "38459": "", "38479": "", "38618": "", "38666": "", "38704": "", "38714": "", "38868": "" }, "source": "meta" }