https://github.com/memgraph/insurance-fraud
Insurance companies lose a lot of money on different kinds of fraud. By using graphs along with machine learning, you can model and detect fraudulent insurance claims.
https://github.com/memgraph/insurance-fraud
Last synced: 10 months ago
JSON representation
Insurance companies lose a lot of money on different kinds of fraud. By using graphs along with machine learning, you can model and detect fraudulent insurance claims.
- Host: GitHub
- URL: https://github.com/memgraph/insurance-fraud
- Owner: memgraph
- License: mit
- Created: 2022-09-01T12:25:21.000Z (over 3 years ago)
- Default Branch: master
- Last Pushed: 2022-09-12T13:08:42.000Z (over 3 years ago)
- Last Synced: 2023-03-03T20:09:46.920Z (about 3 years ago)
- Language: Jupyter Notebook
- Size: 529 KB
- Stars: 2
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
Insurance Fraud Detection Demo
A demo modeling Insurance data, individuals along with their policies, claims, and incidents. Using Memgraph, utilizes the power of graphs to model the data and detect fraudulent claims. Extracts features using graph algorithms and uses machine learning to evaluate insurance claims.
## Data model
## Running the demo
You can set up the project with poetry:
```
poetry install
```
Or, if you don't want to use poetry, simply:
```
pip install -r requirements.txt
```
The demo is located in `./fraud_detection_demo.ipynb`.
Recommended Python version is 3.9, since `sklearn` package doesn't yet support Python 3.10.
## Using the dataset only
If you want to just use the dataset, try running:
```
python dataset/data_generator.py 1000 300
```
And to import it into Memgraph, first start Memgraph, and then run the load script:
```
docker run -it -p 7687:7687 -p 3000:3000 memgraph/memgraph-platform
python load_demo_dataset.py
```