{ "id": 2299, "title": "Getting Started - Python Sample Code (Random Forest)", "url": "/competitions/digit-recognizer/discussion/2299", "author_name": "cclark", "post_date": "2012-08-02T18:12:38.687000", "votes": 28, "comment_count": 10, "views": null, "content": "

Here's some quick Python code to generate the Random Forest benchmark. For help setting up your Python environment and other tips, see https://www.kaggle.com/wiki/GettingStartedWithPythonForDataScience

\r\n

Here's the code - have fun!

\r\n
\r\n
from sklearn.ensemble import RandomForestClassifier
from numpy import genfromtxt, savetxt

def main():
    #create the training & test sets, skipping the header row with [1:]
    dataset = genfromtxt(open('Data/train.csv','r'), delimiter=',', dtype='f8')[1:]    
    target = [x[0] for x in dataset]
    train = [x[1:] for x in dataset]
    test = genfromtxt(open('Data/test.csv','r'), delimiter=',', dtype='f8')[1:]
    
    #create and train the random forest
    #multi-core CPUs can use: rf = RandomForestClassifier(n_estimators=100, n_jobs=2)
    rf = RandomForestClassifier(n_estimators=100)
    rf.fit(train, target)

    savetxt('Data/submission2.csv', rf.predict(test), delimiter=',', fmt='%f')

if __name__=="__main__":
    main()
\r\n
", "messages": [ { "id": 12884, "postDate": "2012-08-02T18:12:38.687Z", "content": "

Here's some quick Python code to generate the Random Forest benchmark. For help setting up your Python environment and other tips, see https://www.kaggle.com/wiki/GettingStartedWithPythonForDataScience

\r\n

Here's the code - have fun!

\r\n
\r\n
from sklearn.ensemble import RandomForestClassifier
from numpy import genfromtxt, savetxt

def main():
    #create the training & test sets, skipping the header row with [1:]
    dataset = genfromtxt(open('Data/train.csv','r'), delimiter=',', dtype='f8')[1:]    
    target = [x[0] for x in dataset]
    train = [x[1:] for x in dataset]
    test = genfromtxt(open('Data/test.csv','r'), delimiter=',', dtype='f8')[1:]
    
    #create and train the random forest
    #multi-core CPUs can use: rf = RandomForestClassifier(n_estimators=100, n_jobs=2)
    rf = RandomForestClassifier(n_estimators=100)
    rf.fit(train, target)

    savetxt('Data/submission2.csv', rf.predict(test), delimiter=',', fmt='%f')

if __name__=="__main__":
    main()
\r\n
", "votes": 28 }, { "id": 135103, "postDate": "2016-09-11T05:23:09.187Z", "content": "

Thanks for the code, but the submission format is a little different from the sample submission file.

\n\n

SO I slightly modify the code for saving csv file as follows:

\n\n
predicted_digits = rf.predict(test)\n\nwith open("rf_benchmark.csv", "wb") as f:\n    writer = csv.writer(f)\n    writer.writerow(["ImageId","Label"])\n    for i, predicted_digit in enumerate(predicted_digits):\n        writer.writerow((i+1, int(predicted_digit)))\n
", "rawMarkdown": "Thanks for the code, but the submission format is a little different from the sample submission file.\r\n\r\nSO I slightly modify the code for saving csv file as follows:\r\n\r\n predicted_digits = rf.predict(test)\r\n\r\n with open(\"rf_benchmark.csv\", \"wb\") as f:\r\n writer = csv.writer(f)\r\n writer.writerow([\"ImageId\",\"Label\"])\r\n for i, predicted_digit in enumerate(predicted_digits):\r\n writer.writerow((i+1, int(predicted_digit)))\r\n", "votes": 2 }, { "id": 140705, "postDate": "2016-10-22T12:06:51.730Z", "content": "

Hi,

\n\n

I couldn't find anything at https://www.kaggle.com/wiki/GettingStartedWithPythonForDataScience

\n\n

can anyone help, please?

", "rawMarkdown": "Hi,\r\n\r\nI couldn't find anything at https://www.kaggle.com/wiki/GettingStartedWithPythonForDataScience\r\n\r\ncan anyone help, please?", "votes": 1 }, { "id": 13725, "postDate": "2012-08-31T23:07:44.907Z", "content": "

To save people just getting starting from having to search for package names:

\r\n

pip install scipy numpy scikit-learn

", "votes": 3 }, { "id": 1534699, "postDate": "2021-10-05T07:00:05.917Z", "content": "

Thank you so much for this short tutorial, this has made the entire process of working with the random forest regressor model very easy. @cclark

", "rawMarkdown": "Thank you so much for this short tutorial, this has made the entire process of working with the random forest regressor model very easy. @cclark " }, { "id": 156940, "postDate": "2017-01-18T17:35:23.053Z", "content": "

Did you try the decode command?\nEg: text.decode(encoding)

", "rawMarkdown": "Did you try the decode command?\r\nEg: text.decode(encoding)" }, { "id": 156937, "postDate": "2017-01-18T17:11:21.890Z", "content": "

hi, \nI'm getting error :\nCan't Convert 'bytes' object to str implicitly\n what should I do ???

", "rawMarkdown": "hi, \r\nI'm getting error :\r\nCan't Convert 'bytes' object to str implicitly\r\n what should I do ???" }, { "id": 72381, "postDate": "2015-04-19T02:37:56.273Z", "content": "

This was super helpful!  Thank you so much!

" }, { "id": 25804, "postDate": "2013-06-17T12:54:13.817Z", "content": "

Hello All.

\r\n

I really appreciate the tutorial in the wiki titled "Getting Started With Python For Data Science". It was signed by Chris Clark and last updated by  Kirstin Rhys. I had no idea where to start on Kaggle and this gave me a starting point.

\r\n

After trying several Python and numerical module installs (I don't get the 2.X versus 3.X series of Python), I finally got around the memory errors and found a combo that would run the Random Forest example (Python 2.7 64-bit windows, with 64-bit windows\r\n versions of NumPy, SciPy, and SciKit-Learn; although I never got setuptools, iPython, or PyDev to work); as long as I didn't try to use more than 1 core.

\r\n

Once I finally got the program to run I got this error: "ValueError: Number of features of the model must  match the input. Model n_features is 784 and  input n_features is 1776".

\r\n

I'll list the whole traceback at the bottom. Any help to get this moving again would be greatly appreciated. I'm stuck and since I was blindly following an example to get a feel for how Kaggle works, I don't know where to begin debugging. The Value Error\r\n seems deep in one of the numerical packages.

\r\n

\r\n

Thanks,

\r\n

-->Rob

\r\n

\r\n

COMPLETE ERROR MESSAGE:

\r\n

Traceback (most recent call last):
\r\nFile "C:\\Users\\Robert\\Documents\\Data Science Class\\week5-Machine Learning\\Kaggle Assignment\\bioresp.py", line 20, in <module>
\r\nmain()
\r\nFile "C:\\Users\\Robert\\Documents\\Data Science Class\\week5-Machine Learning\\Kaggle Assignment\\bioresp.py", line 15, in main
\r\npredicted_probs = [x[1] for x in rf.predict_proba(test)]
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\ensemble\\forest.py", line 560, in predict_proba
\r\nfor i in xrange(n_jobs))
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py", line 514, in __call__
\r\nself.dispatch(function, args, kwargs)
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py", line 311, in dispatch
\r\njob = ImmediateApply(func, args, kwargs)
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py", line 135, in __init__
\r\nself.results = func(*args, **kwargs)
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\ensemble\\forest.py", line 123, in _parallel_predict_proba
\r\nproba_tree = tree.predict_proba(X)
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\tree\\tree.py", line 572, in predict_proba
\r\n% (self.n_features_, n_features))
\r\nValueError: Number of features of the model must match the input. Model n_features is 784 and input n_features is 1776

\r\n

\r\n

\r\n

\r\n

\r\n

\r\n

\r\n

\r\n

\r\n

" }, { "id": 14034, "postDate": "2012-09-07T00:43:18.503Z", "content": "

I seem to be running into MemoryError
\r\nSystem specs:
\r\n4GB RAM
\r\nWindows 7 64 bit
\r\nCore i7 @ 2.00 GHz

" }, { "id": 1619540, "postDate": "2021-12-16T03:18:07.123Z", "content": "

hi. this was super helpful. thank you!

", "rawMarkdown": "hi. this was super helpful. thank you!" } ], "comments": [ { "id": 135103, "author_name": "hysic", "author_url": "", "post_date": "2016-09-11T05:23:09.187000", "content": "

Thanks for the code, but the submission format is a little different from the sample submission file.

\n\n

SO I slightly modify the code for saving csv file as follows:

\n\n
predicted_digits = rf.predict(test)\n\nwith open("rf_benchmark.csv", "wb") as f:\n    writer = csv.writer(f)\n    writer.writerow(["ImageId","Label"])\n    for i, predicted_digit in enumerate(predicted_digits):\n        writer.writerow((i+1, int(predicted_digit)))\n
", "votes": 2, "replies": [] }, { "id": 140705, "author_name": "PRB", "author_url": "", "post_date": "2016-10-22T12:06:51.730000", "content": "

Hi,

\n\n

I couldn't find anything at https://www.kaggle.com/wiki/GettingStartedWithPythonForDataScience

\n\n

can anyone help, please?

", "votes": 1, "replies": [] }, { "id": 13725, "author_name": "ferrouswheel", "author_url": "", "post_date": "2012-08-31T23:07:44.907000", "content": "

To save people just getting starting from having to search for package names:

\r\n

pip install scipy numpy scikit-learn

", "votes": 3, "replies": [] }, { "id": 1534699, "author_name": "HeisenbergOP", "author_url": "", "post_date": "2021-10-05T07:00:05.917000", "content": "

Thank you so much for this short tutorial, this has made the entire process of working with the random forest regressor model very easy. @cclark

", "votes": 0, "replies": [] }, { "id": 156940, "author_name": "Jitesh Krishna", "author_url": "", "post_date": "2017-01-18T17:35:23.053000", "content": "

Did you try the decode command?\nEg: text.decode(encoding)

", "votes": 0, "replies": [] }, { "id": 156937, "author_name": "DaminiKataria", "author_url": "", "post_date": "2017-01-18T17:11:21.890000", "content": "

hi, \nI'm getting error :\nCan't Convert 'bytes' object to str implicitly\n what should I do ???

", "votes": 0, "replies": [] }, { "id": 72381, "author_name": "Kevin Zhao", "author_url": "", "post_date": "2015-04-19T02:37:56.273000", "content": "

This was super helpful!  Thank you so much!

", "votes": 0, "replies": [] }, { "id": 25804, "author_name": "Rob Godbey", "author_url": "", "post_date": "2013-06-17T12:54:13.817000", "content": "

Hello All.

\r\n

I really appreciate the tutorial in the wiki titled "Getting Started With Python For Data Science". It was signed by Chris Clark and last updated by  Kirstin Rhys. I had no idea where to start on Kaggle and this gave me a starting point.

\r\n

After trying several Python and numerical module installs (I don't get the 2.X versus 3.X series of Python), I finally got around the memory errors and found a combo that would run the Random Forest example (Python 2.7 64-bit windows, with 64-bit windows\r\n versions of NumPy, SciPy, and SciKit-Learn; although I never got setuptools, iPython, or PyDev to work); as long as I didn't try to use more than 1 core.

\r\n

Once I finally got the program to run I got this error: "ValueError: Number of features of the model must  match the input. Model n_features is 784 and  input n_features is 1776".

\r\n

I'll list the whole traceback at the bottom. Any help to get this moving again would be greatly appreciated. I'm stuck and since I was blindly following an example to get a feel for how Kaggle works, I don't know where to begin debugging. The Value Error\r\n seems deep in one of the numerical packages.

\r\n

\r\n

Thanks,

\r\n

-->Rob

\r\n

\r\n

COMPLETE ERROR MESSAGE:

\r\n

Traceback (most recent call last):
\r\nFile "C:\\Users\\Robert\\Documents\\Data Science Class\\week5-Machine Learning\\Kaggle Assignment\\bioresp.py", line 20, in <module>
\r\nmain()
\r\nFile "C:\\Users\\Robert\\Documents\\Data Science Class\\week5-Machine Learning\\Kaggle Assignment\\bioresp.py", line 15, in main
\r\npredicted_probs = [x[1] for x in rf.predict_proba(test)]
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\ensemble\\forest.py", line 560, in predict_proba
\r\nfor i in xrange(n_jobs))
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py", line 514, in __call__
\r\nself.dispatch(function, args, kwargs)
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py", line 311, in dispatch
\r\njob = ImmediateApply(func, args, kwargs)
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py", line 135, in __init__
\r\nself.results = func(*args, **kwargs)
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\ensemble\\forest.py", line 123, in _parallel_predict_proba
\r\nproba_tree = tree.predict_proba(X)
\r\nFile "C:\\Python27\\lib\\site-packages\\sklearn\\tree\\tree.py", line 572, in predict_proba
\r\n% (self.n_features_, n_features))
\r\nValueError: Number of features of the model must match the input. Model n_features is 784 and input n_features is 1776

\r\n

\r\n

\r\n

\r\n

\r\n

\r\n

\r\n

\r\n

\r\n

", "votes": 0, "replies": [] }, { "id": 14034, "author_name": "mayank", "author_url": "", "post_date": "2012-09-07T00:43:18.503000", "content": "

I seem to be running into MemoryError
\r\nSystem specs:
\r\n4GB RAM
\r\nWindows 7 64 bit
\r\nCore i7 @ 2.00 GHz

", "votes": 0, "replies": [] }, { "id": 1619540, "author_name": "suptudy", "author_url": "", "post_date": "2021-12-16T03:18:07.123000", "content": "

hi. this was super helpful. thank you!

", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "12884": "", "135103": "Thanks for the code, but the submission format is a little different from the sample submission file.\r\n\r\nSO I slightly modify the code for saving csv file as follows:\r\n\r\n predicted_digits = rf.predict(test)\r\n\r\n with open(\"rf_benchmark.csv\", \"wb\") as f:\r\n writer = csv.writer(f)\r\n writer.writerow([\"ImageId\",\"Label\"])\r\n for i, predicted_digit in enumerate(predicted_digits):\r\n writer.writerow((i+1, int(predicted_digit)))\r\n", "140705": "Hi,\r\n\r\nI couldn't find anything at https://www.kaggle.com/wiki/GettingStartedWithPythonForDataScience\r\n\r\ncan anyone help, please?", "13725": "", "1534699": "Thank you so much for this short tutorial, this has made the entire process of working with the random forest regressor model very easy. @cclark ", "156940": "Did you try the decode command?\r\nEg: text.decode(encoding)", "156937": "hi, \r\nI'm getting error :\r\nCan't Convert 'bytes' object to str implicitly\r\n what should I do ???", "72381": "", "25804": "", "14034": "", "1619540": "hi. this was super helpful. thank you!" } }