{ "id": 10121, "title": "Open .mat file", "url": "/competitions/seizure-prediction/discussion/10121", "author_name": "", "post_date": "2014-08-26T22:42:11.630Z", "votes": 1, "comment_count": 14, "views": 4582, "content": "

Hi guys,

\n

I downloaded one .gz file and I extract it. 

\n

Zip is asking me to name the extension. 

\n

I am using .MAT but I cannot open the unzipped .MAT file later on.

\n

I am using MS Access. Any tips?

\n

Thanks

", "messages": [ { "id": "52481", "postDate": "08/26/2014 22:42:11", "content": "

Hi guys,

\n

I downloaded one .gz file and I extract it. 

\n

Zip is asking me to name the extension. 

\n

I am using .MAT but I cannot open the unzipped .MAT file later on.

\n

I am using MS Access. Any tips?

\n

Thanks

", "rawMarkdown": "", "votes": null }, { "id": "52484", "postDate": "08/27/2014 00:01:53", "content": "

I found the solution in Google as always and NOT in the Kaggle site where it should be.

\n

"Extraction of the XXX.gz file using WinRAR program gave me a huge no-extension file. After renaming the XXX.gz file as XXX.tar.gz, the extraction gave me several .mat files. I hope this "trick" helps you."

", "rawMarkdown": "", "votes": null }, { "id": "52503", "postDate": "08/27/2014 12:16:48", "content": "

Why would you use zip on a gz file?  use gunzip or gzip -d

", "rawMarkdown": "", "votes": null }, { "id": "53039", "postDate": "09/03/2014 15:37:11", "content": "

The files should really be named "Dog_1.tar.gz" etc as they are tarballs (type "file Dog_1.gz" to confirm).  Do the following (the mv isn't necessary, but it helps to see what type of file it is from the name).

\n

  mv Dog_1.gz Dog_1.tar.gz

\n

  tar -xvzf Dog_1.tar.gz

\n

and you'll get files like

\n

Dog_1/Dog_1_interictal_segment_0033.mat
Dog_1/Dog_1_test_segment_0259.mat
Dog_1/Dog_1_test_segment_0292.mat
Dog_1/Dog_1_interictal_segment_0437.mat

", "rawMarkdown": "", "votes": null }, { "id": "53041", "postDate": "09/03/2014 15:39:35", "content": "

Hi,

\n

Yes this is what I did and it worked.

\n

Still cannot open the MAT files though since I do not have MATLAB.  

\n

Hence, I resigned from this competition :)

", "rawMarkdown": "", "votes": null }, { "id": "53043", "postDate": "09/03/2014 15:55:41", "content": "

Use python's scipy.io ("pip install scipy" should do it):

\n\n

http://docs.scipy.org/doc/scipy/reference/generated/scipy.io.loadmat.html#scipy.io.loadmat

\n\n

import scipy.io as sio

\n

mat=sio.loadmat('../epilepsy_data/Dog_1/Dog_1_preictal_segment_0002.mat')

\n\n

mat.keys()
Out[5]: ['preictal_segment_2', '__version__', '__header__', '__globals__']

", "rawMarkdown": "", "votes": null }, { "id": "53052", "postDate": "09/03/2014 17:07:17", "content": "

Thanks for that.

\n

I am using Spyder (Python 3.3) and I wrote:

\n

import scipy.io as sio

\n

mat=sio.loadmat('C:\\Users\\...\\Seizure prediction\\Dog_5\\Dog_5_interictal_segment_0003.mat')

\n

mat.keys()

\n

What I get is: 

\n

SyntaxError: (unicode error) 'unicodeescape' codec can't decode bytes in position 2-3: truncated \\UXXXXXXXX escape

", "rawMarkdown": "", "votes": null }, { "id": "53053", "postDate": "09/03/2014 17:13:43", "content": "

position 2-3 of "c:\\U" is in your string.  Try putting the .mat file in your current directory and opening it.

", "rawMarkdown": "", "votes": null }, { "id": "53072", "postDate": "09/03/2014 20:46:16", "content": "

You can use 'gnu octave' its free and does almost the same as matlab

", "rawMarkdown": "", "votes": null }, { "id": "53701", "postDate": "09/14/2014 15:50:34", "content": "

You can use R to open a .mat file.  You first have to download R, which is an open source program for data analysis and visualization.  You then have to install the R.matlab package, also free, which allows you to read and write .mat files.  Below is the website for the R homepage.  Hope this helps.

\n

http://www.r-project.org/

", "rawMarkdown": "", "votes": null }, { "id": "53804", "postDate": "09/15/2014 15:38:43", "content": "

The problem with the "c:\\U" is the "\\" character.

\n\n

Some systems/languages to represent complex characters use combinations of "\\" and another letter. For example "\\t" is tab and "\\n" means "new line"

\n\n

Try using the character "/" instead of "\\" or alternatively use the complex representation for "\\" which is "\\\\".

\n

Summing it up. Try one of the following:

\n

"C:/Users/..."

\n

"C:\\\\Users\\\\..."

", "rawMarkdown": "", "votes": null }, { "id": "57145", "postDate": "11/01/2014 00:48:19", "content": "

Hello Omid

\n

I have downloaded and opened the file in R as you suggested. However I cannot seem to access the variable. as it is too large.

\n

How did you manipulate the data (say for dog_1 the smallest) after you downloaded the data in R?

\n

I admit I usually work on the theory side of this and now I am trying to perform more hands on.

\n\n

Thanks

\n\n

v/r

\n\n

-Ray

", "rawMarkdown": "", "votes": null }, { "id": "57146", "postDate": "11/01/2014 01:06:56", "content": "

Apologies this was already solved.

\n

Thanks

\n

v/r

\n

-Ray

", "rawMarkdown": "", "votes": null }, { "id": "58102", "postDate": "11/15/2014 07:15:17", "content": "

I am new to matlab format. Anybody can give me some hint how to go from mat['preictal_segment_2'] to data/data_length_sec/sampling_frequency/channels/sequence ? 

\n

I found mat['preictal_segment_2'] is 1*1 numpy.ndarray, but when i go to mat['preictal_segment_2'][0][0], I got a numpy.void element.

\n\n\n\n

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

\n

>>Use python's scipy.io ("pip install scipy" should do it):

\n

>> http://docs.scipy.org/doc/scipy/reference/generated/scipy.io.loadmat.html#scipy.io.loadmat

\n

>> import scipy.io as sio

\n

>> mat=sio.loadmat('../epilepsy_data/Dog_1/Dog_1_preictal_segment_0002.mat')

\n

>> mat.keys()
>> Out[5]: ['preictal_segment_2', '__version__', '__header__', '__globals__']

", "rawMarkdown": "", "votes": null }, { "id": "58103", "postDate": "11/15/2014 07:46:01", "content": "

no need to reply. I found the answer myself -- 

\n\n

mat['preictal_segment_2'][0][0] can be further traced down to four elements mat['preictal_segment_2'][0][0][0] ~mat['preictal_segment_2'][0][0][3]

", "rawMarkdown": "", "votes": null } ], "comments": [ { "id": 52484, "author_name": "kchlou", "author_url": "", "post_date": "08/27/2014 00:01:53", "content": "

I found the solution in Google as always and NOT in the Kaggle site where it should be.

\n

"Extraction of the XXX.gz file using WinRAR program gave me a huge no-extension file. After renaming the XXX.gz file as XXX.tar.gz, the extraction gave me several .mat files. I hope this "trick" helps you."

", "votes": null, "replies": [] }, { "id": 52503, "author_name": "galexander", "author_url": "", "post_date": "08/27/2014 12:16:48", "content": "

Why would you use zip on a gz file?  use gunzip or gzip -d

", "votes": null, "replies": [] }, { "id": 53039, "author_name": "cwilkes", "author_url": "", "post_date": "09/03/2014 15:37:11", "content": "

The files should really be named "Dog_1.tar.gz" etc as they are tarballs (type "file Dog_1.gz" to confirm).  Do the following (the mv isn't necessary, but it helps to see what type of file it is from the name).

\n

  mv Dog_1.gz Dog_1.tar.gz

\n

  tar -xvzf Dog_1.tar.gz

\n

and you'll get files like

\n

Dog_1/Dog_1_interictal_segment_0033.mat
Dog_1/Dog_1_test_segment_0259.mat
Dog_1/Dog_1_test_segment_0292.mat
Dog_1/Dog_1_interictal_segment_0437.mat

", "votes": null, "replies": [] }, { "id": 53041, "author_name": "kchlou", "author_url": "", "post_date": "09/03/2014 15:39:35", "content": "

Hi,

\n

Yes this is what I did and it worked.

\n

Still cannot open the MAT files though since I do not have MATLAB.  

\n

Hence, I resigned from this competition :)

", "votes": null, "replies": [] }, { "id": 53043, "author_name": "cwilkes", "author_url": "", "post_date": "09/03/2014 15:55:41", "content": "

Use python's scipy.io ("pip install scipy" should do it):

\n\n

http://docs.scipy.org/doc/scipy/reference/generated/scipy.io.loadmat.html#scipy.io.loadmat

\n\n

import scipy.io as sio

\n

mat=sio.loadmat('../epilepsy_data/Dog_1/Dog_1_preictal_segment_0002.mat')

\n\n

mat.keys()
Out[5]: ['preictal_segment_2', '__version__', '__header__', '__globals__']

", "votes": null, "replies": [] }, { "id": 53052, "author_name": "kchlou", "author_url": "", "post_date": "09/03/2014 17:07:17", "content": "

Thanks for that.

\n

I am using Spyder (Python 3.3) and I wrote:

\n

import scipy.io as sio

\n

mat=sio.loadmat('C:\\Users\\...\\Seizure prediction\\Dog_5\\Dog_5_interictal_segment_0003.mat')

\n

mat.keys()

\n

What I get is: 

\n

SyntaxError: (unicode error) 'unicodeescape' codec can't decode bytes in position 2-3: truncated \\UXXXXXXXX escape

", "votes": null, "replies": [] }, { "id": 53053, "author_name": "cwilkes", "author_url": "", "post_date": "09/03/2014 17:13:43", "content": "

position 2-3 of "c:\\U" is in your string.  Try putting the .mat file in your current directory and opening it.

", "votes": null, "replies": [] }, { "id": 53072, "author_name": "ruiprodrigues", "author_url": "", "post_date": "09/03/2014 20:46:16", "content": "

You can use 'gnu octave' its free and does almost the same as matlab

", "votes": null, "replies": [] }, { "id": 53701, "author_name": "omidrnasserbigdeli", "author_url": "", "post_date": "09/14/2014 15:50:34", "content": "

You can use R to open a .mat file.  You first have to download R, which is an open source program for data analysis and visualization.  You then have to install the R.matlab package, also free, which allows you to read and write .mat files.  Below is the website for the R homepage.  Hope this helps.

\n

http://www.r-project.org/

", "votes": null, "replies": [] }, { "id": 53804, "author_name": "jmarinero", "author_url": "", "post_date": "09/15/2014 15:38:43", "content": "

The problem with the "c:\\U" is the "\\" character.

\n\n

Some systems/languages to represent complex characters use combinations of "\\" and another letter. For example "\\t" is tab and "\\n" means "new line"

\n\n

Try using the character "/" instead of "\\" or alternatively use the complex representation for "\\" which is "\\\\".

\n

Summing it up. Try one of the following:

\n

"C:/Users/..."

\n

"C:\\\\Users\\\\..."

", "votes": null, "replies": [] }, { "id": 57145, "author_name": "stormer1120", "author_url": "", "post_date": "11/01/2014 00:48:19", "content": "

Hello Omid

\n

I have downloaded and opened the file in R as you suggested. However I cannot seem to access the variable. as it is too large.

\n

How did you manipulate the data (say for dog_1 the smallest) after you downloaded the data in R?

\n

I admit I usually work on the theory side of this and now I am trying to perform more hands on.

\n\n

Thanks

\n\n

v/r

\n\n

-Ray

", "votes": null, "replies": [] }, { "id": 57146, "author_name": "stormer1120", "author_url": "", "post_date": "11/01/2014 01:06:56", "content": "

Apologies this was already solved.

\n

Thanks

\n

v/r

\n

-Ray

", "votes": null, "replies": [] }, { "id": 58102, "author_name": "orchid", "author_url": "", "post_date": "11/15/2014 07:15:17", "content": "

I am new to matlab format. Anybody can give me some hint how to go from mat['preictal_segment_2'] to data/data_length_sec/sampling_frequency/channels/sequence ? 

\n

I found mat['preictal_segment_2'] is 1*1 numpy.ndarray, but when i go to mat['preictal_segment_2'][0][0], I got a numpy.void element.

\n\n\n\n

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

\n

>>Use python's scipy.io ("pip install scipy" should do it):

\n

>> http://docs.scipy.org/doc/scipy/reference/generated/scipy.io.loadmat.html#scipy.io.loadmat

\n

>> import scipy.io as sio

\n

>> mat=sio.loadmat('../epilepsy_data/Dog_1/Dog_1_preictal_segment_0002.mat')

\n

>> mat.keys()
>> Out[5]: ['preictal_segment_2', '__version__', '__header__', '__globals__']

", "votes": null, "replies": [] }, { "id": 58103, "author_name": "orchid", "author_url": "", "post_date": "11/15/2014 07:46:01", "content": "

no need to reply. I found the answer myself -- 

\n\n

mat['preictal_segment_2'][0][0] can be further traced down to four elements mat['preictal_segment_2'][0][0][0] ~mat['preictal_segment_2'][0][0][3]

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "52481": "", "52484": "", "52503": "", "53039": "", "53041": "", "53043": "", "53052": "", "53053": "", "53072": "", "53701": "", "53804": "", "57145": "", "57146": "", "58102": "", "58103": "" }, "source": "meta" }