Experiments in malware detection and classification using machine learning techniques.
https://www.kaggle.com/c/malware-classification
Initial feature engineering consisted of extracting various keyword counts from the ASM files
as well as the entropy and file size from the BYTE files of the 10868 malware samples in the training set.
Image files of the first 1000 bytes of the ASM and BYTE files were created and combined with
keyword and entropy data. This resulted in a set of 2018 features.
Flow control graphs and call graphs were generated for each ASM sample. A feature set was
then generated from the graphs, including graph maximum delta, density, diameter and function
counts etc.
Statistical analysis of the feature set using chi-squared tests to remove features that are
independent of the class labels or have low variance. The BYTE file images were found to be weak
learners and were removed from the feature set. A comparison of the best features from the chi-squared
tests with reduced feature sets of between 10% - 50% of the original features.
Testing with an ExtraTreesClassifier and 10-fold cross validation produced the following results:
- Original ASM Keyword Counts (1006 features): logloss = 0.034
- 10% Best ASM Features with Entropy and Image Features (202 features): logloss = 0.0174
- 20% Best ASM with Entropy and Image Features (402 features): logloss = 0.0164
- 30% Best ASM with Entropy and Image Features plus Feature Statistics (623 features):
multiclass logloss = 0.0133
accuracy score = 0.9978
Confusion Matrix:
[[1540 0 0 0 0 1 0 0 0]
[ 1 2475 2 0 0 0 0 0 0]
[ 0 0 2942 0 0 0 0 0 0]
[ 1 0 0 474 0 0 0 0 0]
[ 2 0 0 0 38 2 0 0 0]
[ 3 0 0 0 0 748 0 0 0]
[ 1 0 0 0 0 0 397 0 0]
[ 0 0 0 0 0 0 0 1225 3]
[ 0 0 0 0 0 0 0 8 1005]]
- 40% Best ASM and image features with feature statistics:
ExtraTreesClassifier with 1000 estimators on 10868 training samples and 823 features
using 10-fold cross validation:
multiclass logloss = 0.0135
accuracy score = 0.9976
Confustion Matrix:
[[1541 0 0 0 0 0 0 0 0]
[ 1 2475 2 0 0 0 0 0 0]
[ 0 0 2942 0 0 0 0 0 0]
[ 1 0 0 474 0 0 0 0 0]
[ 5 0 0 0 37 0 0 0 0]
[ 5 0 0 0 0 746 0 0 0]
[ 1 0 0 0 0 0 397 0 0]
[ 0 0 0 0 0 0 0 1227 1]
[ 0 0 0 0 0 0 0 9 1004]]
The performance of the ExtraTreesClassifier is optimal at around 30% of ASM and image features
with highest variance plus sample statistics, entropy and file size. Adding call graph features
produced a marginal improvement. It is possible that better classification accuracy would be
achieved by using an ensemble of different classifiers with the ASM, image and call graph
feature sets as separate inputs to the various classifiers.
Selection of candidate models using GridSearchCV to find optimal classifier hyper-parameters.
- SVM:
- ExtraTrees:
- XGBoost: 30% Best Features
logloss: 0.0080
accuracy: 0.9981
Confusion Matrix:
[[1540 0 0 0 0 1 0 0 0]
[ 2 2475 0 1 0 0 0 0 0]
[ 0 0 2941 0 0 0 1 0 0]
[ 0 0 0 474 0 1 0 0 0]
[ 1 0 0 0 41 0 0 0 0]
[ 4 0 0 0 1 746 0 0 0]
[ 0 0 0 0 0 0 398 0 0]
[ 0 0 0 0 0 0 0 1227 1]
[ 0 0 0 0 0 0 0 8 1005]]
- NaiveBayes:
- KNN:
1. Shannon's Entropy by malware class. A score of 0.0 means the bytes are all the same value,
a score of 1.0 means every byte in the file has a different value.
2. Shannon's Entropy by file size. A score of 0.0 means the bytes are all the same value,
a score of 1.0 means every byte in the file has a different value.
3. Assembler register EDX by ESI counts.
The best accuracy scores were achieved with XGBoost (99.81%) and ExtraTreesClassifier (99.76%) using a feature set of 623 ASM, image and entropy features. Marginal improvements could be achieved using additional features and ensemble methods, however due to the limited sample size further efforts are unlikely to produce significant improvements in prediction accuracy. Analysis will now focus on much larger sample sizes from VirusShare.com as described in the following sections.
VirusShare.com regularly publishes huge collections of malware binaries for use by researchers. Each malware archive is currently around 25GB in size. Several of the latest archives have been downloaded to use as training and test sets. The archives used are:
- Training set: VirusShare_00251.zip and VirusShare_00252.zip (131072 malware samples)
VirusShare_00263.zip and VirusShare_00264.zip (131072 malware samples)
VirusShare_APT1_293.zip (293 malware samples)
- Testing set:
Using Cuckoo Sandbox and unpack.py for behaviourial analysis, unpacking the binaries and
dumping process memory, for intransigent samples, manual unpacking with Immunity Debugger and IDA Pro.
Tools:
- Cuckoo Sandbox (https://github.com/cuckoosandbox/cuckoo)
- unpack.py (https://malwaremusings.com/2013/02/26/automated-unpacking-a-behaviour-based-approach/)
(https://github.com/malwaremusings/unpacker)
- IDA Pro 5.0 (https://www.hex-rays.com/products/ida/support/download_freeware.shtml)
- Immunity Debugger (https://www.immunityinc.com/products/debugger/)
- Volatility (https://github.com/volatilityfoundation)
- Ildasm.exe (https://msdn.microsoft.com/en-us/library/f7dy01k1(v=vs.110).aspx)
- ndisasm (http://www.nasm.us/pub/nasm/releasebuilds/2.12.02/)
- TrID (http://mark0.net/soft-trid-e.html)
- ClamAV (clamav.net)
- Windows Defender
- MalwareBytes Anti-Malware
- VirusTotal.com
apt install virtualbox virtualbox-dkms python-dev libffi-dev virtualenv virtualenvwrapper clamav
pip install cython numpy scipy scikit-learn matplotlib jupyter pandas xgboost
git clone https://github.com/cuckoosandbox/cuckoo
git clone https://github.com/volatility
TODO:
ClamAV and Windows Defender used for initial training label generation or VirusTotal.com aggregate classification
if they cannot identify the culprit. MalwareBytes was also used but it crashed at the end of the scan
and the log files could not be recovered.
AV Scan Results:
Results: VirusShare_00251
- 57529 files classified as malicious.
- 8007 files classified as non-malicious.
Results: VirusShare_00252
- 56625 files classified as malicious.
- 8911 files classified as non-malicious.
Results: VirusShare_00263
- 51612 files classified as malicious.
- 13924 files classified as non-malicious.
Results: VirusShare_00264
- 42274 files classified as malicious.
- 23262 files classified as non-malicious.
Results: VirusShare_APT1_293
- 292 files classified as malicious.
- 1 file classified as non-malicious.
Total Malware Types: 8334
Total Malware Families: 2737
Total Files: 262437
4. Top 10 Malware Counts.
5. Top 10 Compiler/Packer Counts.
6. VirusShare 251 Call Graph - Vertex by Edge Count.
7. VirusShare 251 Shannon's File Entropy Histogram.
IDA Pro and objdump for disassembly of binaries to .asm text files.
Feature sets will consist of:
- Entropy and file size from packed binaries.
- Entropy and file size from unpacked binaries.
- File magic signatures and TrID signatures.
- ASM features from disassembled unpacked binaries.
- Executable header features.
- Call Graph Features.
- Function counts extracted from call graphs.
- Sample Statistics.
- Behavioural features from Cuckoo Sandbox reports.
- Memory features from Volatility reports.
1. PE/COFF Binaries: (Chi2 Tests)
VS251 Feature Sets: 54911 samples.
240 PE ASM and Header Features.
?? PE ASM Function Count Features.
VS252 Feature Sets: 46165 samples.
271 PE ASM and Header Features.
?? PE ASM Function Count Features.
VS263 Feature Sets: 40974 samples.
203 PE ASM and Header Features.
?? PE ASM Function Count Features.
VS264 Feature Sets: 14366 samples.
243 PE ASM and Header Features.
?? PE ASM Function Count Features.
2. ELF Binaries:
3. Java Bytecode:
4. Javascript:
5. HTML:
6. PDF:
Model selection with 10-fold cross validation:
1. ExtraTreesClassifier: VS251 100 estimators accuracy score = 0.912
500 estimators accuracy score = ?.??
1000 estimators accuracy score = memory fail
VS252 100 estimators accuracy score = 0.888 (12.75 minutes)
500 estimators accuracy score = ?.??
1000 estimators accuracy score = ?.??
VS263 100 estimators accuracy score = 0.903 (9.63 minutes)
500 estimators accuracy score = ?.???
1000 estimators accuracy score = ?.??
VS264 100 estimators accuracy score = 0.889 (2.27 minutes)
500 estimators accuracy score = 0.890 (14.57 minutes)
1000 estimators accuracy score = ?.??
2. XGBoost: VS251 100 estimators accuracy score = ?.??
XGBoost: VS252 100 estimators accuracy score = ?.??
XGBoost: VS263 100 estimators accuracy score = ?.??
XGBoost: VS264 100 estimators accuracy score = ?.??
3. LightGBM: VS251 100 estimators accuracy score = 0.892
VS252 100 estimators accuracy score = 0.676 (171.23 minutes)
VS263 100 estimators accuracy score = ?.??
VS264 100 estimators accuracy score = 0.758 (9.26 minutes)
200 estimators accuracy score = 0.750 (18.53 minutes)
4. RandomForestClassifier: VS251 100 estimators accuracy score = 0.903
500 estimators accuracy score = ?.??
1000 estimators accuracy score = ?.??
VS252 100 estimators accuracy score = 0.881 (81.34 minutes)
VS263 100 estimators accuracy score = ?.??
VS264 100 estimators accuracy score = 0.879 (15.45 minutes)
Model Stacks/Ensembles:
1. One input layer of classifiers -> 1 output layer classifier.
Layer 1: Six x layer one classifiers: (ExtraTrees x 2, RandomForest x 2, XGBoost x 1, LightGBM x 1)
Layer 2: One classifier: (ExtraTrees) -> final labels
2. Voting (Democratic and weighted).
Democratic: Six x layer one classifiers: (ExtraTrees x 2/RandomForest x 2/XGBoost/LightGBM)
-> (democratic vote, geometric and sum means) -> final labels
Weighted: Six x layer one classifiers: (ExtraTrees x 2/RandomForest x 2/XGBoost/LightGBM)
-> (weighted vote: ExtraTrees double weight, geometric and sum means) -> final labels
3. Multiple layers of classifiers.
Layer one -> layer two -> layer 3 -> final labels:
Layer 1: ExtraTrees x 2, RandomForest x 2, XGBoost x 1, LightGBM x 1
Layer 2: ExtraTrees x 2, RandomForest x 2, XGBoost x 1, LightGBM x 1
Layer 3: ExtraTrees x 1
4. Combined PE/COFF features + function count features:
Layer 1 -> layer 2 -> final labels
Layer 1 (A MODELS): Combined features layer one (ExtraTrees x 2, RandomForest x 2, XGBoost x 1, LightGBM x 1)
Layer 1 (B MODELS): Function count features layer one (ExtraTrees x 2, RandomForest x 2, XGBoost x 1, LightGBM x 1)
Layer 2: ExtraTrees x 1 -> final labels
5. Combine outputs from 1, 2, 3 and 4 -> vote -> final labels
TODO:
TODO:
1. Antivirus scans using ClamAV and Windows Defender.
> clamscan -v -r /directory/containing/the/nastiness > clamav-report.txt
> Windows Defender (See notes in section 7 on extracting windows defender logs).
2. Generate scalar training labels for each malware type and family.
> process_av_reports.py
> combine_av_reports.py
> generate_train_labels.py
1. File entropy feature generation.
> feature_extraction_entropy.py
2. File magic signature and TrID signature feature generation.
> trid_check_file.py
> generate_file_ids.py
> feature_extraction_file_id.py
3. Packer identification feature generation.
> generate_packer_ids.py
> feature_extraction_packer_id.py
4. ASM feature generation (unpacked PE files).
> disassemble_pe.py
> feature_extraction_pe_asm.py
> generate_pe_header_tokens.py
> feature_extraction_pe_header.py
5. ASM feature generation (packed PE files).
> TODO:
6. Call Graph Generation and feature extraction.
> generate_call_graphs_pe_asm.py
> generate_function_column_names.py
> function_name_clean.py
> feature_extraction_pe_function_counts.py
> feature_reduction_pe_function_counts.py
7. Behavioural analysis feature generation.
> TODO:
8. Memory analysis feature generation.
> TODO:
1. File entropy feature generation.
> feature_extraction_entropy.py
2. File magic signature and TrID signature feature generation.
> trid_check_file.py
> generate_file_ids.py
> feature_extraction_file_id.py
3. Packer identification feature generation.
> generate_packer_ids.py
> feature_extraction_packer_id.py
4. ASM feature generation.
> disassemble_elf.py
> feature_extraction_elf_asm.py
5. Call Graph Generation.
>
6. Behavioural analysis feature generation.
>
7. Memory analysis feature generation.
>
1. Convert Bytecode to Tokens.
2. Extract Bytecode Features.
Tools:
javap (https://docs.oracle.com/javase/7/docs/technotes/tools/windows/javap.html)
1. Generate Javascript/HTML Keywords.
2. Unpack Javascript.
3. Extract Javascript/HTML Features.
Tools:
1. Generate PDF Keywords.
2. Extract Javascript/Shellcode/Macros.
3. Extract PDF Feature Sets.
Tools:
peepdf (https://github.com/jesparza/peepdf)
1. PE/COFF Feature Reduction.
> feature_reduction_pe_asm.py
> feature_reduction_pe_header.py
> feature_reduction_pe_function_counts.py
TODO:
1. PE/COFF Model Selection.
> model_selection_pe_coff.py
TODO:
TODO:
If installing from source, after building and installing you have problems loading other packages it is because of the xgboost-0.4-py2.7.egg.pth file that the install script dumps in the python dist-packages directory. You will have to delete the .pth file then go change the installation of the xgboost egg and egg-info files in the python dist-packages directory from:
/usr/local/lib/python2.7/dist-packages/xgboost-0.4-py2.7.egg/EGG_INFO
to:
/usr/local/lib/python2.7/dist-packages/xgboost-0.4-py2.7.dist-info
and:
/usr/local/lib/python2.7/dist-packages/xgboost-0.4-py2.7.egg/xgboost
to:
/usr/local/lib/python2.7/dist-packages/xgboost
Now python will be able to find all the packages.
pip install xgboost
Now works for version 0.6a2 on Debian/Ubuntu/Mint distros.
XGBoost is not a part of the official distribution but several community members have created Conda packages for it. The most up to date package seems to be by user creditx. The following command will install the package:
conda install -c creditx xgboost
Python 2.7 is preferred for Cuckoo Sandbox, attempting with Python 3.x will be a fail. Installing the Python module requirements in requirements.txt results in failure because the module dpkt is only compatible with Python 2.x versions. If using Anaconda or python 3.x then revert to Python 2.7 or use mkvirtualenv to create a virtual environment to run cuckoo.
For example:
mkvirtualenv -p /usr/bin/python cuckoosandbox
Note: If using Anaconda: Remove the Anaconda bin directory from $PATH or it will cause an error when setting up the virtual environment. Also ensure that libxml2-dev and libxslt1-dev are installed or there will be build errors when installing the requirements.
Open a powershell - (Run as Administrator).
Enter the following commands:
> cd \Program Files\Windows Defender
> .\MpCmdRun -getfiles -scan
Several .log and .cab files will be placed in:
C:\ProgramData\Microsoft\Windows Defender\Support\
The Windows Defender malware detection log is called MPDetection-yymmdd-hhmm.log
Ensure binutils multi-target support has been installed (Linux Mint 18):
(NOTE: Linux Mint 17 does not have MIPS architecture in binutils, have to install from sauce.)
apt install binutils binutils-aarch64-linux-gnu binutils-alpha-linux-gnu binutils-arm-linux-gnueabi
binutils-arm-linux-gnueabihf binutils-arm-linux-gnueabihf binutils-arm-none-eabi binutils-avr binutils-dev
binutils-doc binutils-gold binutils-h8300-hms binutils-hppa-linux-gnu binutils-hppa64 binutils-hppa64-linux-gnu
binutils-m68hc1x binutils-m68k-linux-gnu binutils-mingw-w64 binutils-mingw-w64-i686 binutils-mingw-w64-x86-64
binutils-mips-linux-gnu binutils-mips64-linux-gnuabi64 binutils-mips64-linux-gnuabi64 binutils-mips64el-linux-gnuabi64
binutils-mips64el-linux-gnuabi64 binutils-mipsel-linux-gnu binutils-msp430 binutils-multiarch binutils-multiarch-dev
binutils-powerpc-linux-gnu binutils-powerpc-linux-gnuspe binutils-powerpc-linux-gnuspe binutils-powerpc64-linux-gnu
binutils-powerpc64-linux-gnu binutils-powerpc64le-linux-gnu binutils-powerpc64le-linux-gnu binutils-s390x-linux-gnu
binutils-sh4-linux-gnu binutils-source binutils-sparc64-linux-gnu binutils-z80 elf-binutils
1. Clone, build and install:
git clone --recursive https://github.com/Microsoft/LightGBM
cd LightGBM
mkdir build
cd build
cmake ..
make -j
cd ../python-package/
python setup.py install
2. If you have problems with building or installing python module:
apt update
apt upgrade
apt install cmake
pip install setuptools numpy scipy scikit-learn -U
3. If you have problems with updating setuptools, sklearn etc, and you probably will because (pip == train wreck):
apt purge -y python-pip
wget https://bootstrap.pypa.io/get-pip.py
python ./get-pip.py
apt install python-pip
pip install setuptools numpy scipy scikit-learn -U