Dataset generating from APK testing
Budget: $30 – $250 CAD
Instructions
Key Points to cover :
Implementation of any testing tool with apk that gives provided features and the tools support the idea of detecting malicious activity on the application.
Develop a Python program that performs actions like - install apk, test and feature extract and record in the dataset.
Details Instruction( Bold mark to carefully work)
Please download the repo from Github (that I will provide )and there is a folder with 100 apk. These will be our testing sample.
There are a few resources we provided - the framework of tool and research paper if you need them.
First, find out the android application testing tool that gives similar features as in the template dataset. And send us a log file sample that you are getting most API call signature, manifest and all.
( There are 215 features that won't be in all apps but try to cover as many as possible. At least after analyzing close to 200 features because it's a traditional testing result, features should have all in most applications.)
Note : (Try to develop the program- in GitHub folder -so you can push back to the server, after all, is done, use of visual code and android studio emulator recommended and Linux environment)
Now, important to develop a complete python program that :
Install 1 apk at a time from the selected APK folder
(Suggestion: when we try to install multiple apk usually a program need to be put inside android SDK and put the folder path to install - But do your own way)
Then- the program performs the test with that installed apk.
(If you wish, you can store the log file too)
Then the python program - count the number of occurrences/ frequency of those 215 features called while testing.
Doing feature extraction, parsing those numbers from log or direct -store in the dataset. Like: how many times “ x” called or “y’ called or “z” called – total number of times those features from 215 appears in the testing time.
In the dataset- Top column names would feature, and the first left column would be the package name and in all rows number of times. ( see the provided - Dataset-Template) .
Important:
Run test for 30 sec
Run the test for 100 apk- there might few not get installed - we will give extra and also does not matter little more or less.
Make the program in a way that - Next time, if I try to test more- I just need to provide another folder name, and it does the same performance. For your first 100Apk - APkfolder 1, then for my next 100, I will create a folder apk2, then I can use the program.
Push all programs to GitHub it's possible.
Provide us details notes- how to run what; try to use python3 and provide us with any emulator API requirement. We are very open to that.
Key Points to cover :
Implementation of any testing tool with apk that gives provided features and the tools support the idea of detecting malicious activity on the application.
Develop a Python program that performs actions like - install apk, test and feature extract and record in the dataset.
Details Instruction( Bold mark to carefully work)
Please download the repo from Github (that I will provide )and there is a folder with 100 apk. These will be our testing sample.
There are a few resources we provided - the framework of tool and research paper if you need them.
First, find out the android application testing tool that gives similar features as in the template dataset. And send us a log file sample that you are getting most API call signature, manifest and all.
( There are 215 features that won't be in all apps but try to cover as many as possible. At least after analyzing close to 200 features because it's a traditional testing result, features should have all in most applications.)
Note : (Try to develop the program- in GitHub folder -so you can push back to the server, after all, is done, use of visual code and android studio emulator recommended and Linux environment)
Now, important to develop a complete python program that :
Install 1 apk at a time from the selected APK folder
(Suggestion: when we try to install multiple apk usually a program need to be put inside android SDK and put the folder path to install - But do your own way)
Then- the program performs the test with that installed apk.
(If you wish, you can store the log file too)
Then the python program - count the number of occurrences/ frequency of those 215 features called while testing.
Doing feature extraction, parsing those numbers from log or direct -store in the dataset. Like: how many times “ x” called or “y’ called or “z” called – total number of times those features from 215 appears in the testing time.
In the dataset- Top column names would feature, and the first left column would be the package name and in all rows number of times. ( see the provided - Dataset-Template) .
Important:
Run test for 30 sec
Run the test for 100 apk- there might few not get installed - we will give extra and also does not matter little more or less.
Make the program in a way that - Next time, if I try to test more- I just need to provide another folder name, and it does the same performance. For your first 100Apk - APkfolder 1, then for my next 100, I will create a folder apk2, then I can use the program.
Push all programs to GitHub it's possible.
Provide us details notes- how to run what; try to use python3 and provide us with any emulator API requirement. We are very open to that.