{ "id": 12505, "title": "Feature Extraction from Training Data", "url": "/competitions/malware-classification/discussion/12505", "author_name": "", "post_date": "2015-02-13T15:09:58.617Z", "votes": 1, "comment_count": 4, "views": 2782, "content": "

Hi,

\n

I am new to data science and machine learning a complete novice. I took part in this competition to learn practical applications of ML. When I went through the training data I found it is made up of ".asm" files, any ideas on how I can extract features from these files. I am a newbie in the field.

\n

 

\n

Also is this the right approach?

\n

Thanks in advance. :)

\n

 

\n

P.S.: Good-luck everyone.

", "messages": [ { "id": "64165", "postDate": "02/13/2015 15:09:58", "content": "

Hi,

\n

I am new to data science and machine learning a complete novice. I took part in this competition to learn practical applications of ML. When I went through the training data I found it is made up of ".asm" files, any ideas on how I can extract features from these files. I am a newbie in the field.

\n

 

\n

Also is this the right approach?

\n

Thanks in advance. :)

\n

 

\n

P.S.: Good-luck everyone.

", "rawMarkdown": "", "votes": null }, { "id": "64166", "postDate": "02/13/2015 15:27:34", "content": "

Hi guys even I m a learner, I saw in datasample.7z that we were provided .asm and .bytes but in training.7z there are no .bytes files, so now do we have to use "hexdump" to convert the .asm to .bytes?

", "rawMarkdown": "", "votes": null }, { "id": "64167", "postDate": "02/13/2015 15:37:59", "content": "

[quote=eNVy;64166]

\n

Hi guys even I m a learner, I saw in datasample.7z that we were provided .asm and .bytes but in training.7z there are no .bytes files, so now do we have to use "hexdump" to convert the .asm to .bytes?

\n

[/quote]

\n

there are both asm and bytes files in training.7z. Either your extraction failed or there was some download error.

", "rawMarkdown": "", "votes": null }, { "id": "64168", "postDate": "02/13/2015 15:38:33", "content": "

[quote=Shanky Sharma;64165]

\n

Hi,

\n

I am new to data science and machine learning a complete novice. I took part in this competition to learn practical applications of ML. When I went through the training data I found it is made up of ".asm" files, any ideas on how I can extract features from these files. I am a newbie in the field.

\n

 

\n

Also is this the right approach?

\n

Thanks in advance. :)

\n

 

\n

P.S.: Good-luck everyone.

\n

[/quote]

\n\n

Take a look at the benchmark post in the forums. Although it deals with the bytes files, it will give you some idea.

", "rawMarkdown": "", "votes": null }, { "id": "64169", "postDate": "02/13/2015 15:38:43", "content": "

[quote=eNVy;64166]

\n

Hi guys even I m a learner, I saw in datasample.7z that we were provided .asm and .bytes but in training.7z there are no .bytes files, so now do we have to use "hexdump" to convert the .asm to .bytes?

\n

[/quote]

\n

Are you sure your train.7z is not corrupt (check md5s)? There are .bytes files in my archive.

\n

[quote=Shanky Sharma;64165]

\n

Hi,

\n

I am new to data science and machine learning a complete novice. I took part in this competition to learn practical applications of ML. When I went through the training data I found it is made up of ".asm" files, any ideas on how I can extract features from these files. I am a newbie in the field.

\n

Also is this the right approach?

\n

Thanks in advance. :)

\n

P.S.: Good-luck everyone.

\n

[/quote]

\n

Feature extraction is a very important (if not the single most important) part of this contest, so people are not likely to share nonbasic approaches, but you can still check "beat the benchmark" and "tutorial" forum posts.

\n

If you want more sophisticated approaches, you can for example treat the .asm files as text - there's a lot of tutorials and papers on extracting features from text, just google it. It may or may not prove to be effective in this competition.

\n

PS. Good luck to you too.

\n

PS2. Are you sure you've chosen the right competition to start your data science adventure? The size of the data makes verifying your ideas quite time-consuming.

", "rawMarkdown": "", "votes": null } ], "comments": [ { "id": 64166, "author_name": "neerajvash8", "author_url": "", "post_date": "02/13/2015 15:27:34", "content": "

Hi guys even I m a learner, I saw in datasample.7z that we were provided .asm and .bytes but in training.7z there are no .bytes files, so now do we have to use "hexdump" to convert the .asm to .bytes?

", "votes": null, "replies": [] }, { "id": 64167, "author_name": "abhishek", "author_url": "", "post_date": "02/13/2015 15:37:59", "content": "

[quote=eNVy;64166]

\n

Hi guys even I m a learner, I saw in datasample.7z that we were provided .asm and .bytes but in training.7z there are no .bytes files, so now do we have to use "hexdump" to convert the .asm to .bytes?

\n

[/quote]

\n

there are both asm and bytes files in training.7z. Either your extraction failed or there was some download error.

", "votes": null, "replies": [] }, { "id": 64168, "author_name": "abhishek", "author_url": "", "post_date": "02/13/2015 15:38:33", "content": "

[quote=Shanky Sharma;64165]

\n

Hi,

\n

I am new to data science and machine learning a complete novice. I took part in this competition to learn practical applications of ML. When I went through the training data I found it is made up of ".asm" files, any ideas on how I can extract features from these files. I am a newbie in the field.

\n

 

\n

Also is this the right approach?

\n

Thanks in advance. :)

\n

 

\n

P.S.: Good-luck everyone.

\n

[/quote]

\n\n

Take a look at the benchmark post in the forums. Although it deals with the bytes files, it will give you some idea.

", "votes": null, "replies": [] }, { "id": 64169, "author_name": "datarat", "author_url": "", "post_date": "02/13/2015 15:38:43", "content": "

[quote=eNVy;64166]

\n

Hi guys even I m a learner, I saw in datasample.7z that we were provided .asm and .bytes but in training.7z there are no .bytes files, so now do we have to use "hexdump" to convert the .asm to .bytes?

\n

[/quote]

\n

Are you sure your train.7z is not corrupt (check md5s)? There are .bytes files in my archive.

\n

[quote=Shanky Sharma;64165]

\n

Hi,

\n

I am new to data science and machine learning a complete novice. I took part in this competition to learn practical applications of ML. When I went through the training data I found it is made up of ".asm" files, any ideas on how I can extract features from these files. I am a newbie in the field.

\n

Also is this the right approach?

\n

Thanks in advance. :)

\n

P.S.: Good-luck everyone.

\n

[/quote]

\n

Feature extraction is a very important (if not the single most important) part of this contest, so people are not likely to share nonbasic approaches, but you can still check "beat the benchmark" and "tutorial" forum posts.

\n

If you want more sophisticated approaches, you can for example treat the .asm files as text - there's a lot of tutorials and papers on extracting features from text, just google it. It may or may not prove to be effective in this competition.

\n

PS. Good luck to you too.

\n

PS2. Are you sure you've chosen the right competition to start your data science adventure? The size of the data makes verifying your ideas quite time-consuming.

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "64165": "", "64166": "", "64167": "", "64168": "", "64169": "" }, "source": "meta" }