Data Stream Clustering Algorithms

This is an implementation of some well known Data Stream Clustering Algorithms for my engineering thesis @ PWR Algorithmic Computer Science. This project also includes sampling wrappers for aforementioned algorithms.

Author

Maksymilian Neumann

Requirements

1. Internet connection
2. Rust toolchain
3. Python 3.12
4. Install python libraries

pip install -r requirements.txt

5. Generate data

python gen_all_data.py

Usage

1. Run

cargo run -r

2. Choose any benchmarks or run all
3. Plot to see results in ./plots

python gen_all_plots.py

4. (Optional) generate demo results for k-means and DBSCAN as a comparison

python classic_demo.py

Implementations

Algorithms:

• BIRCH
• CluStream
• DenStream

Samplers:

• Static
• Dynamic

Data Used

Real data used:

B. S. and R. Nagapadma. "RT-IoT2022 ," UCI Machine Learning Repository, 2023. [Online]. Available: https://doi.org/10.24432/C5P338.