The premise of this project is to examine the C functions used by web browsers and see what information can be extrapolated from them. Variations on this concept include running based on limited information (small windows of functions).

This repo contains the standard program that reads an entire strace file as well as the program that reads small chunks of strace files and determines which site they're generated from.

Note: This repo is similar to motif-site_grouping, but has more output files available. Mostly for record keeping purposes.

#Requirements#

• SciPy
• NumPy
• Scikit-learn

#Program Notes#

curl/wget:

www.google.com www.reddit.com www.yahoo.com www.microsoft.com www.wfu.edu www.nyu.edu www.netflix.com www.amazon.com www.soundcloud.com www.youtube.com

//TODO websites that don't exist set some flags downloading files big websites (lots of traffic)

ls:

~/ ~/Documents ~/Desktop ~/Downloads ~/grive ~/Music ~/Pictures ~/Public ~/Videos ~/Templates

//TODO add flags (for all dirs done before)

• done, used flags -a, -h, -l, -F nonexistent directory

strace:

ls wget www.google.com curl www.google.com cal dir df ping www.google.com vi date dc

more/less: output_parser_0.3.py window_classifier.py curl_1.txt curl_10.txt ls_1.txt ls_10.txt strace_1.txt strace_10.txt wget_1.txt wget_10.txt

new ones:

ssh telnet sftp scp ftp

can you predict an application will end? vector for end lines vector for elsewhere id's for each, run and see

checking number of connections:

ps -ef more proc/[proc_num]/net/sockstat

wget www.google.com &

Goal:

create sliding window (n, n+1, etc.) that serves as testing set use current output chunks as testing set see if utility can make a prediction