Explain the logic

[licaon-kter]

...of the thing that classifies the apps, what input does it get? APKs? Permissions? etc

[sanandmv7]

The app extracts permissions and intents of installed applications from the user's device. These extracted features are then fed to the machine learning model in the form of a vector. The machine learning model returns a prediction score between 0 and 1 that denote the degree of maliciousness of the scanned application. We use this score to classify the scanned app into one of the following categories:

1. Goodware: The prediction score is less than 0.5
2. Risky: Prediction score between 0.5 and 0.75
3. Malware: Prediction score is greater than 0.75
4. Unknown: If our app could not extract permissions and intents from any of the installed apps, then those apps are labelled as 'Unknown'

You can check the code for building machine learning model here

[licaon-kter]

Do add this info on the site or Readme :)