Skip to content

P7XCKD/File-type-prediction

Repository files navigation

Smart File Type Prediction Problem: Just checking a file’s name or extension isn’t safe, because someone can rename a harmful file (like a program .exe) to look like a picture .jpg.

ML Solution The model looks inside the file’s raw data (its first bytes), learns the hidden patterns of each file type, and predicts what the file really is. If that doesn’t match the given extension, it’s flagged as suspicious.


Smart File Type Prediction: Project Plan

Overview

This project builds a machine learning system to predict the true type of a file by analyzing its raw byte data, helping to flag suspicious files that may have misleading extensions.


Project Structure

test_folder/
├── original/          # Contains files with correct extensions
│   ├── document.pdf
│   ├── image.jpg
│   ├── music.mp3
│   ├── program.exe
│   └── ...
├── modified/          # Contains same files with wrong extensions (auto-generated)
│   ├── document.jpg   # PDF file disguised as JPG
│   ├── image.exe      # JPG file disguised as EXE
│   └── ...
└── models/            # Trained ML models and data
    ├── file_classifier.pkl
    └── training_data.csv

]

Expected Output

=== File Type Prediction Results ===
✓ document.pdf → Predicted: PDF, Extension: PDF (SAFE)
⚠ document.jpg → Predicted: PDF, Extension: JPG (SUSPICIOUS!)
✓ image.png → Predicted: PNG, Extension: PNG (SAFE)
⚠ image.exe → Predicted: PNG, Extension: EXE (SUSPICIOUS!)

Detection Accuracy: 95%
Suspicious Files Found: 2/4

About

mlt project

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages