# Deep Learning for Cyber Security — Data

来源：https://www.kaggle.com/competitions/deep-learning-for-cyber-security/data

采集时间：2026-10-20T02:52:53Z

The dataset represents a group of APK files (rows) and their privileges (columns).

It contains **8078** samples in total.
4039 are Malicious and 4039 are Non-Malicious.
Split into 70% for train.csv and 30% for test.csv.


## Files

*   **train.csv** - the training set (contains all features and classes)
*   **test.csv** - the test set (contains all features without the class) 
* **sample_submission** - an example of how your submission file should look like


## Columns

*   `id` - unique numerical value for each APK file
*   `Malicious` - identify the class of the APK file (FALSE is benign and TRUE is malicious)
*   The dataset contains 70 columns which identify the privileges of the APK file