Streaming Data Structures Analysis

Overview

This project explores probabilistic and streaming algorithms for analyzing clickstream data under memory constraints. It compares approximate methods against exact results to study the trade-offs between efficiency, scalability, and accuracy.

Methods

• Bloom Filter
• Count-Min Sketch
• Flajolet–Martin
• DGIM sliding window approximation

Data

Clickstream Data for Online Shopping
UCI Machine Learning Repository

Tools

Python, NumPy, pandas, matplotlib, seaborn, mmh3