node2vec

A production-ready three-tier anime recommendation engine combining SvelteKit (frontend), NestJS (API bridge), and Rust (ML engine) with PostgreSQL, Neo4j, and Redis.

Future research collaboration prediction and partner recommendation using link prediction methods and Node2Vec on scientific co-author networks. (Dự đoán hợp tác nghiên cứu tương lai và gợi ý đối tác khoa học bằng các thuật toán dự đoán liên kết và Node2Vec trên mạng đồng tác giả.)

Co-Occurance Vertex Embedding

Zero-to-hero for Graph Neural Networks

Graph-Tabular Fusion for Bitcoin Fraud Detection - Demonstrating when Node2Vec embeddings don't improve XGBoost. Scientifically rigorous negative result validating that tabular features encode graph structure.

Solutions to Stanford CS224w Machine Learning with Graphs assignments — shallow node embeddings and message-passing GNNs (GCN/GraphSAGE/GAT) for node/graph classification and link prediction with PyTorch Geometric, with measured accuracies

Aim is to convert nodes and node attributes of the DBLP Citation graph to analyze graph specific trends. This objective entailed two tasks, recreating a search algorithm for sampling the neighborhood as per the Node2Vec algorithm and extract feature embeddings using the Word2Vec skip-gram architecture. The nodes (papers)are represented into a fi…

GraphSAGE link-prediction prototype exploring candidate airline routes across 3,500+ airports using network and geographic features.

ArXiv co-author recommendation system based on link predictions

This repository contains the code and results of my Machine Learning on Graphs course project, where I applied various community detection methods to the OGBN-ArXiv dataset.

🔗 Explore Graph-Tabular Fusion to enhance Bitcoin fraud detection, revealing that rich tabular features render graph embeddings redundant.

Embed music albums into a latent space leveraging their reviews by Pitchfork.

Node2Vec 2nd-order biased random walk generator with return (p) and in-out (q) hyperparameters

This project implements a hybrid recommendation system for Amazon Video Games by modeling product relationships through a co-review graph. It combines behavioral signals, structural graph features, and textual content to predict relevant "Related Items" for a cross-sell widget.

Interactive Marimo WASM app for teaching spectral, node2vec, and GraphSAGE node embeddings with node classification.

한국 청년 임업인 산촌 진입 의사결정 지원 시스템

Node2Vec 2nd-order biased random walk generator with return (p) and in-out (q) hyperparameters

Graph neural network for fraud detection: GNN on transaction graphs, node2vec embeddings, community detection for fraud rings, real-time scoring API and explainable alerts

ATLAS is a classic word game where players name geographical locations. Each name must start with the last letter of the previous one. The game ends when a player cannot provide a valid, previously unused location. This project analyzes the geographical word game ATLAS using complex network theory.

Final assignment of EE226 course in SJTU by Group 12