# sohu9/zero-day-fraud-detection architecture

> Architecture diagram of the GitHub repository sohu9/zero-day-fraud-detection, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-02). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/sohu9/zero-day-fraud-detection
- Repository: https://github.com/sohu9/zero-day-fraud-detection

## Overview

This repository is a zero-day payment-fraud detection application: it learns normal transaction behavior with an autoencoder, flags anomalous transactions, and presents risk decisions with explanations. The main workflow runs from transaction submission through the Flask application’s preprocessing and inference to a decision, persistence, and dashboard/API access. Separate training and evaluation workflows prepare the model and measure its performance. The README also describes SHAP/LIME interpretability and broader behavioral features; sampled code shows feature-based explanation text, but does not establish SHAP/LIME execution. Some application wiring is omitted from the excerpts.

Main entry points: [`app.py`](https://github.com/sohu9/zero-day-fraud-detection/blob/main/app.py) (Flask transaction service), [`evaluate.py`](https://github.com/sohu9/zero-day-fraud-detection/blob/main/evaluate.py) (Offline evaluation), [`train.py`](https://github.com/sohu9/zero-day-fraud-detection/blob/main/train.py) (Model training), [`data_pipeline.py`](https://github.com/sohu9/zero-day-fraud-detection/blob/main/data_pipeline.py) (Feature preparation)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_modeling["Data and modeling"]
  node_data_pipeline["Feature preparation<br/>[data_pipeline.py]"]
  node_preprocessor["Fitted preprocessor<br/>[data_pipeline.py]"]
  node_trainer["Model training<br/>[train.py]"]
  node_model_builder["Autoencoder builder<br/>[model_builder.py]"]
  node_model["Trained autoencoder<br/>[train.py]"]
  node_threshold["Anomaly threshold<br/>[train.py]"]
end

subgraph group_serving["Fraud decision service"]
  node_web_app["Flask transaction service<br/>[app.py]"]
  node_live_preprocessing["Live preprocessing<br/>[app.py]"]
  node_inference["Anomaly scoring<br/>[app.py]"]
  node_explanation["Decision explanation<br/>[app.py]"]
  node_dashboard["Web dashboard<br/>[index.html]"]
end

subgraph group_records["Transaction records"]
  node_transaction_db[("Transaction store<br/>[database.py]")]
end

subgraph group_assessment["Model assessment"]
  node_evaluation["Offline evaluation<br/>[evaluate.py]"]
  node_metrics["Evaluation results<br/>[evaluate.py]"]
end

node_customer(("Transaction submitter"))
node_training_data["Training transactions"]
node_test_data["Test transactions"]

node_training_data -->|"loads and transforms"| node_data_pipeline
node_trainer -->|"runs preprocessing"| node_data_pipeline
node_data_pipeline -->|"saves fitted state"| node_preprocessor
node_trainer -->|"builds model"| node_model_builder
node_trainer -->|"trains and saves"| node_model
node_trainer -->|"calibrates and saves"| node_threshold
node_customer -.->|"submits transaction"| node_web_app
node_web_app -->|"prepares input"| node_live_preprocessing
node_web_app -->|"analyzes transaction"| node_inference
node_web_app -->|"builds explanation"| node_explanation
node_web_app -->|"queries records"| node_transaction_db
node_web_app -.->|"serves web interface"| node_dashboard
node_test_data -->|"evaluates against"| node_evaluation
node_evaluation -->|"uses time features"| node_data_pipeline
node_evaluation -->|"loads for scoring"| node_model
node_evaluation -->|"loads threshold"| node_threshold
node_evaluation -->|"writes results"| node_metrics

click node_data_pipeline "https://github.com/sohu9/zero-day-fraud-detection/blob/main/data_pipeline.py"
click node_preprocessor "https://github.com/sohu9/zero-day-fraud-detection/blob/main/data_pipeline.py"
click node_trainer "https://github.com/sohu9/zero-day-fraud-detection/blob/main/train.py"
click node_model_builder "https://github.com/sohu9/zero-day-fraud-detection/blob/main/model_builder.py"
click node_model "https://github.com/sohu9/zero-day-fraud-detection/blob/main/train.py"
click node_threshold "https://github.com/sohu9/zero-day-fraud-detection/blob/main/train.py"
click node_web_app "https://github.com/sohu9/zero-day-fraud-detection/blob/main/app.py"
click node_live_preprocessing "https://github.com/sohu9/zero-day-fraud-detection/blob/main/app.py"
click node_inference "https://github.com/sohu9/zero-day-fraud-detection/blob/main/app.py"
click node_explanation "https://github.com/sohu9/zero-day-fraud-detection/blob/main/app.py"
click node_dashboard "https://github.com/sohu9/zero-day-fraud-detection/blob/main/templates/index.html"
click node_transaction_db "https://github.com/sohu9/zero-day-fraud-detection/blob/main/database.py"
click node_evaluation "https://github.com/sohu9/zero-day-fraud-detection/blob/main/evaluate.py"
click node_metrics "https://github.com/sohu9/zero-day-fraud-detection/blob/main/evaluate.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_data_pipeline,node_preprocessor,node_trainer,node_model_builder,node_model,node_threshold toneBlue
class node_web_app,node_live_preprocessing,node_inference,node_explanation,node_dashboard toneAmber
class node_transaction_db toneMint
class node_evaluation,node_metrics toneRose
class node_customer,node_training_data,node_test_data toneIndigo
```

## Components

### Data and modeling

- [Feature preparation](https://github.com/sohu9/zero-day-fraud-detection/blob/main/data_pipeline.py): `data_pipeline.py`
- [Fitted preprocessor](https://github.com/sohu9/zero-day-fraud-detection/blob/main/data_pipeline.py): `data_pipeline.py`
- [Model training](https://github.com/sohu9/zero-day-fraud-detection/blob/main/train.py): `train.py`
- [Autoencoder builder](https://github.com/sohu9/zero-day-fraud-detection/blob/main/model_builder.py): `model_builder.py`
- [Trained autoencoder](https://github.com/sohu9/zero-day-fraud-detection/blob/main/train.py): `train.py`
- [Anomaly threshold](https://github.com/sohu9/zero-day-fraud-detection/blob/main/train.py): `train.py`

### Fraud decision service

- [Flask transaction service](https://github.com/sohu9/zero-day-fraud-detection/blob/main/app.py): `app.py`
- [Live preprocessing](https://github.com/sohu9/zero-day-fraud-detection/blob/main/app.py): `app.py`
- [Anomaly scoring](https://github.com/sohu9/zero-day-fraud-detection/blob/main/app.py): `app.py`
- [Decision explanation](https://github.com/sohu9/zero-day-fraud-detection/blob/main/app.py): `app.py`
- [Web dashboard](https://github.com/sohu9/zero-day-fraud-detection/blob/main/templates/index.html): `templates/index.html`

### Transaction records

- [Transaction store](https://github.com/sohu9/zero-day-fraud-detection/blob/main/database.py): `database.py`

### Model assessment

- [Offline evaluation](https://github.com/sohu9/zero-day-fraud-detection/blob/main/evaluate.py): `evaluate.py`
- [Evaluation results](https://github.com/sohu9/zero-day-fraud-detection/blob/main/evaluate.py): `evaluate.py`

### Other components

- **Transaction submitter**
- **Training transactions**
- **Test transactions**

## Connections

- Training transactions → Feature preparation: loads and transforms (evidence: `data_pipeline.py`)
- Model training → Feature preparation: runs preprocessing (evidence: `train.py`)
- Feature preparation → Fitted preprocessor: saves fitted state (evidence: `data_pipeline.py`)
- Model training → Autoencoder builder: builds model (evidence: `train.py`)
- Model training → Trained autoencoder: trains and saves (evidence: `train.py`)
- Model training → Anomaly threshold: calibrates and saves (evidence: `train.py`)
- Transaction submitter → Flask transaction service: submits transaction (evidence: `README.md`)
- Flask transaction service → Live preprocessing: prepares input (evidence: `app.py`)
- Flask transaction service → Anomaly scoring: analyzes transaction (evidence: `app.py`)
- Flask transaction service → Decision explanation: builds explanation (evidence: `app.py`)
- Flask transaction service → Transaction store: queries records (evidence: `app.py`)
- Flask transaction service → Web dashboard: serves web interface (evidence: `README.md`)
- Test transactions → Offline evaluation: evaluates against (evidence: `evaluate.py`)
- Offline evaluation → Feature preparation: uses time features (evidence: `evaluate.py`)
- Offline evaluation → Trained autoencoder: loads for scoring (evidence: `evaluate.py`)
- Offline evaluation → Anomaly threshold: loads threshold (evidence: `evaluate.py`)
- Offline evaluation → Evaluation results: writes results (evidence: `evaluate.py`)
