Graph Recommendation

A project focused on constructing a graph from raw data and developing a recommendation system leveraging the graph structure. The system combines traditional algorithms with neural network-based methods to provide recommendations.

Dataset: https://grouplens.org/datasets/movielens/

---

to run:

1. git clone https://github.com/xxesth/cGraphRecommendation.git
2. cd cGraphRecommendation
3. download dataset, extract and move it to the project directory
4. make
5. ./build/run

to debug:

1. gcc -g -o debug src/main.c src/graph.c src/recommendation.c src/neuralnetwork.c -lm
2. gdb ./debug
3. (gdb) run
4. (gdb) backtrace or (gdb) print X

---

.
|-- LICENSE
|-- Makefile 
|-- README.md  
|-- .gitignore
|-- projectdescription.pdf
|-- projectreport.pdf
|-- ml-100k # data
|   |-- u.data  # main dataset
|-- model
|   |-- trained_model.bin
|-- src  
|   |-- graph.c  
|   |-- graph.h  
|   |-- main.c  
|   |-- neuralnetwork.c  
|   |-- neuralnetwork.h  
|   |-- recommendation.c  
|   |-- recommendation.h

---

notes:

1. Matrix factorization neural network is implemented for this project. GNN and NCF are other possible solutions.
2. Content based recommendation can be included as a further step of this project.