# ๐ Africa Banking Crisis Prediction
### ๐ง Deep Learning ร Economic Intelligence
Predicting banking-crisis conditions from historical African economic indicators using Deep Learning.
Overview โข
Architecture โข
Model โข
Evaluation โข
Deployment
---
## ๐ From Economic Data to Crisis Intelligence
Financial crises rarely emerge from a single indicator.
They are shaped by a complex interaction of **inflation, currency instability, sovereign debt, domestic debt, systemic stress and broader economic conditions**.
This project explores whether those historical patterns can be learned by a **Deep Neural Network** and transformed into an interactive prediction system.
> **Can historical economic signals help us identify banking-crisis conditions?**
This project attempts to answer that question using a complete **Data โ Deep Learning โ Evaluation โ Deployment** pipeline.
---
## โก What This Project Does
```text
๐ ECONOMIC DATA
โ
โผ
๐ DATA ANALYSIS
โ
โผ
๐งน PREPROCESSING
โ
โโโโโโโโดโโโโโโโ
โผ โผ
Numerical Categorical
Scaling Encoding
โ โ
โโโโโโโโฌโโโโโโโ
โผ
๐ง DEEP NEURAL
NETWORK
โ
โผ
๐ MODEL TRAINING
โ
โผ
๐ฏ EVALUATION
โ
โผ
๐พ SAVED MODEL
โ
โผ
๐ STREAMLIT APP
โ
โผ
๐ฎ CRISIS PREDICTION
```
---
## ๐ง The Core Idea
The system takes historical economic and financial indicators as input and learns relationships between those indicators and the `banking_crisis` target.
### Input
```text
๐ฆ Banking Indicators
๐ฐ Debt Indicators
๐ Inflation Indicators
๐ฑ Currency Indicators
๐ Country Information
๐
Historical Information
โ ๏ธ Crisis Indicators
```
โฌ๏ธ
### Deep Learning Model
```text
Input Features
โ
Dense Layer โ 128 Neurons
โ
ReLU Activation
โ
Dropout โ 30%
โ
Dense Layer โ 64 Neurons
โ
ReLU Activation
โ
Dropout โ 30%
โ
Softmax Output
```
โฌ๏ธ
### Output
```text
๐ข NO CRISIS
OR
๐ด BANKING CRISIS
```
---
## ๐ Project at a Glance
| ๐ Component | ๐ ๏ธ Implementation |
| -------------------- | --------- โฆ