Update README.md

README.md

@@ -1,420 +1,338 @@
 ---
-title: stockpredictor
-sdk: gradio
-emoji: 😻
+title: StockPredictor
+emoji: 📈
 colorFrom: blue
 colorTo: indigo
-sdk_version: 6.14.0
+sdk: gradio
+sdk_version: "5.29.0"
+app_file: app.py
 pinned: false
+license: mit
 ---
-# 📈 Stock Price Predictor
+
+---
+
+title: StockPredictor emoji: 📈 colorFrom: blue colorTo: indigo sdk: gradio sdk_version: "5.29.0" app_file: app.py pinned: false license: mit
+
+📈 Stock Price Predictor
 
 AI-powered stock price predictions using LSTM neural networks, technical analysis, and market sentiment analysis.
 
-**Features:**
-- ⚡ Fast predictions (<2 seconds)
-- 🧠 Deep Learning LSTM model
-- 📊 Technical indicators (RSI, MACD, Volatility, etc.)
-- 💭 News sentiment analysis
-- 💾 24-hour prediction cache
-- 🎯 85-86% directional accuracy
-- 📱 Web interface (Gradio)
+🚀 Features
+
+⚡ Fast predictions
+
+🧠 LSTM Deep Learning model
+
+📊 Technical indicators (RSI, MACD, SMA, ROC)
+
+💭 News sentiment analysis
+
+💾 Prediction caching
+
+📱 Gradio web interface
+
+🌍 Supports US and Indian stocks
+
+
 
 ---
 
-## 🚀 Quick Start (Local)
+🛠️ Local Setup
 
-### 1. Clone or Download Files
+1. Clone Repository
 
-```bash
-# Download all files from this repository
-# You should have:
-# - app.py
-# - model.py
-# - requirements.txt
-# - .env.example
-# - .gitignore
-# - README.md
-```
+git clone https://github.com/your-username/stock-predictor.git
+cd stock-predictor
+
+2. Create Virtual Environment
 
-### 2. Create Virtual Environment (Recommended)
+Windows
 
-```bash
-# Windows
 python -m venv venv
 venv\Scripts\activate
 
-# macOS/Linux
+macOS/Linux
+
 python3 -m venv venv
 source venv/bin/activate
-```
 
-### 3. Install Dependencies
 
-```bash
+---
+
+3. Install Dependencies
+
 pip install -r requirements.txt
-```
 
-### 4. Setup Environment Variables
 
-```bash
-# Copy example file to .env
-cp .env.example .env
+---
+
+4. Add Environment Variables
+
+Create a .env file:
 
-# Edit .env and add your API key:
-# FINNHUB_API_KEY=your_api_key_here
-```
+FINNHUB_API_KEY=your_api_key_here
 
-**Get Free API Key:**
-- Go to https://finnhub.io/register
-- Sign up (free tier available)
-- Copy your API key to `.env`
+Get a free API key from:
+
+https://finnhub.io/register
+
+
+
+---
 
-### 5. Run the App
+5. Run Application
 
-```bash
 python app.py
-```
 
-The app will start at `http://localhost:7860`
+The application will start at:
+
+http://localhost:7860
+
 
 ---
 
-## 🌐 Deploy on HuggingFace Spaces
+🌐 Deploy on Hugging Face Spaces
 
-### Step 1: Create HF Space
+Step 1: Create Space
 
-1. Go to https://huggingface.co/spaces
-2. Click "Create new Space"
-3. Fill in:
-   - **Space name**: `stock-predictor` (or any name)
-   - **License**: `mit` or `apache-2.0`
-   - **SDK**: Select **Gradio**
-   - **Visibility**: Public or Private
+1. Open:
 
-4. Click "Create Space"
 
-### Step 2: Upload Files
 
-You'll see a command to clone your space. Then:
+https://huggingface.co/spaces
+
+2. Click:
+
+
+
+Create New Space
+
+3. Configure:
+
 
-```bash
-cd your-space-repo
-git clone https://huggingface.co/spaces/YOUR_USERNAME/stock-predictor
-cd stock-predictor
-```
 
-Copy these files into the directory:
-```
+SDK: Gradio
+
+License: MIT
+
+Visibility: Public or Private
+
+
+
+---
+
+Step 2: Upload Project Files
+
+Upload the following files:
+
 app.py
 model.py
 requirements.txt
+README.md
 .env.example
 .gitignore
-README.md
-```
 
-### Step 3: Setup Environment Variables (HF Spaces)
 
-In HF Spaces, go to **Settings** → **Repository secrets** → **Add secret**
+---
+
+Step 3: Add Repository Secrets
 
-Add:
-- **Name**: `FINNHUB_API_KEY`
-- **Value**: Your Finnhub API key
+Go to:
 
-**DO NOT** commit `.env` file to git. HF Spaces will automatically load secrets as environment variables.
+Settings → Repository secrets
 
-### Step 4: Push to HF Spaces
+Add the following secret:
 
-```bash
-git add .
-git commit -m "Initial stock predictor setup"
-git push
-```
+Name: FINNHUB_API_KEY
+Value: your_api_key
 
-HF Spaces will automatically:
-1. Install dependencies from `requirements.txt`
-2. Run `app.py`
-3. Deploy your app with a public URL
 
 ---
 
-## 📖 Usage Guide
+Step 4: Automatic Deployment
 
-### Via Web Interface
+Hugging Face Spaces will automatically:
 
-1. Enter stock symbol (e.g., `AAPL`, `GOOGL`, `MSFT`)
-2. Select number of days to predict (1-30)
-3. Click **🚀 Predict**
-4. View results table and analysis
+Install dependencies
 
-### Supported Symbols
+Run app.py
 
-Any stock ticker from:
-- **US**: NASDAQ, NYSE (AAPL, GOOGL, MSFT, TSLA, etc.)
-- **Europe**: LSE, Euronext (various symbols)
-- **India**: NSE, BSE (SBIN.NS, INFY.NS, etc.)
+Create a public deployment URL
 
-### Example Stocks
 
-```
-AAPL    - Apple
-GOOGL   - Google (Alphabet)
-MSFT    - Microsoft
-AMZN    - Amazon
-TSLA    - Tesla
-NVDA    - NVIDIA
-META    - Meta (Facebook)
-NFLX    - Netflix
-```
 
 ---
 
-## 🔧 How It Works
-
-### 1. Data Collection (Step 1)
-- Fetches 100 days of historical price data from Yahoo Finance
-- Uses: Open, High, Low, Close, Volume prices
-
-### 2. Feature Engineering (Step 2)
-Calculates 6 technical indicators:
-- **Close**: Stock closing price
-- **RSI**: Relative Strength Index (momentum, 14-day)
-- **MACD**: Moving Average Convergence Divergence
-- **Volatility**: Standard deviation of returns (10-day)
-- **SMA20**: 20-day Simple Moving Average (trend)
-- **ROC**: Rate of Change (5-day)
-
-### 3. Preprocessing (Step 3)
-- Normalizes all features to [0, 1] range using MinMaxScaler
-- Creates sequences of 30 days for LSTM training
-
-### 4. LSTM Model Training (Step 4)
-- **Architecture**: 1 LSTM layer (32 hidden units) + 1 Dense layer
-- **Training**: 50 epochs, batch size 32
-- **Optimizer**: Adam (lr=0.001)
-- **Loss**: Mean Squared Error
-- **Time**: ~2 minutes on CPU
-
-### 5. Price Prediction (Step 5)
-- Uses trained LSTM to predict next N days
-- Autoregressive approach: uses previous predictions for future predictions
-- Inverse transforms predictions back to actual prices
-
-### 6. Sentiment Analysis (Step 6)
-- Fetches latest 10 news articles from Finnhub
-- Uses VADER sentiment analyzer (rule-based, fast, accurate)
-- Classifies each article as Positive/Negative/Neutral
+📊 Supported Stock Symbols
 
----
+US Stocks
 
-## ⚡ Performance
+AAPL
+MSFT
+GOOGL
+AMZN
+TSLA
+NVDA
+META
+NFLX
 
-| Metric | Value |
-|--------|-------|
-| **Single Prediction** | 1-2 seconds |
-| **Cached Prediction** | 50-100ms |
-| **Model Training** | 2 minutes |
-| **Memory Usage** | 2-3GB |
-| **Package Size** | 600MB |
-| **Accuracy (MAPE)** | 4.8% |
-| **Directional Accuracy** | 76-77% |
+Indian Stocks
 
-### Optimization Techniques
+SBIN.NS
+INFY.NS
+RELIANCE.NS
+TCS.NS
+HDFCBANK.NS
 
-- ✅ PyTorch instead of TensorFlow (6x smaller)
-- ✅ yfinance instead of Alpha Vantage (no API limits)
-- ✅ Direct LSTM (no XGBoost residuals)
-- ✅ VADER sentiment (no transformers)
-- ✅ 24-hour prediction cache
-- ✅ Gradio native support
 
 ---
 
-## 📊 Model Architecture
+🧠 Model Architecture
+
+Input Features
+      ↓
+LSTM Layer (32 Units)
+      ↓
+Dense Layer
+      ↓
+Predicted Price
 
-```
-Input Sequence (30 days × 6 features)
-           ↓
-      LSTM Layer
-    (32 hidden units)
-           ↓
-      Dense Layer
-       (1 output)
-           ↓
-   Predicted Price
-```
 
 ---
 
-## 🎯 Accuracy Analysis
+📈 Technical Indicators Used
 
-### Why 85-86%?
+RSI
 
-Stock predictions face inherent limitations:
-- **Market efficiency**: ~95% of info already priced in
-- **Black swan events**: Unpredictable shocks
-- **Noise**: High-frequency trading noise
+MACD
 
-### MAPE (Mean Absolute Percentage Error)
+SMA20
 
-- **Interpretation**: Average prediction error
-- **Value**: 4.8% (±5% of stock price)
-- **Trade-off**: Acceptable for directional trading
+Volatility
+
+ROC
+
+Closing Price
 
-### Directional Accuracy
 
-- **Definition**: Predicting if price goes UP or DOWN
-- **Accuracy**: 76-77%
-- **Better than**: Random guess (50%)
 
 ---
 
-## 🚨 Limitations & Disclaimers
+⚡ Performance
+
+Metric	Value
+
+Prediction Speed	1-2 seconds
+Cached Predictions	50-100ms
+Training Time	~2 minutes
+Memory Usage	2-3 GB
+Accuracy (MAPE)	~4.8%
+Directional Accuracy	76-77%
 
-⚠️ **This is for educational purposes only. NOT financial advice.**
 
-### Limitations
 
-1. **Historical data dependent**: Assumes past patterns repeat
-2. **No fundamental analysis**: Ignores earnings, dividends, etc.
-3. **No geopolitical events**: Can't predict wars, pandemics
-4. **Short-term only**: Best for 1-7 day predictions
-5. **Sentiment limited**: Only recent news, not long-term sentiment
+---
+
+⚠️ Disclaimer
+
+This project is for educational purposes only.
 
-### What This Can't Predict
+Not financial advice
 
-- Company earnings surprises
-- FDA drug approvals
-- Geopolitical events
-- Market crashes
-- Sector rotations
+Markets are unpredictable
 
-### Investment Disclaimer
+Always conduct your own research
+
+Invest responsibly
 
-**Past performance ≠ Future results**
 
-Always:
-- ✅ Conduct your own research
-- ✅ Diversify your portfolio
-- ✅ Consult a financial advisor
-- ✅ Manage risk properly
-- ✅ Never invest what you can't afford to lose
 
 ---
 
-## 🔐 Security
+🔐 Security Notes
 
-### API Keys
+Never upload .env
 
-- **Never commit** `.env` file
-- Use **HF Spaces Secrets** for deployment
-- API keys are not logged or stored
-- Data comes from Yahoo Finance (public)
+Use HF Spaces Secrets for API keys
 
-### Data Privacy
+No user data stored
 
-- ✅ No user data stored
-- ✅ Predictions cached locally (24 hours)
-- ✅ No ads or tracking
-- ✅ Open source code
+Uses public Yahoo Finance data
 
----
 
-## 🐛 Troubleshooting
 
-### "ModuleNotFoundError: No module named 'torch'"
+---
 
-```bash
-pip install -r requirements.txt
-```
+🐛 Common Issues
+
+Configuration Error on HF Spaces
 
-### "No data for symbol XXXX"
+Use this exact YAML block at the top of README.md:
 
-1. Check symbol spelling (e.g., `AAPL` not `APPLE`)
-2. Use international format (e.g., `SBIN.NS` for Indian stocks)
-3. Check stock is listed on a major exchange
 
-### "FINNHUB_API_KEY not set"
 
-Sentiment analysis will be skipped, but predictions will still work:
-1. Get key: https://finnhub.io/register
-2. Add to `.env` file
-3. Restart app
+Important:
 
-### "CUDA out of memory"
+YAML must start at line 1
 
-Will automatically fall back to CPU (slower but works)
+No spaces before ---
 
-### First prediction is slow
+app_file is required
 
-This is expected (model training takes ~2 minutes). Subsequent predictions use cached model (<2 seconds).
+Use supported Gradio versions only
 
-### "Connection timeout" or "Too many requests"
 
-API rate limit reached. Solution:
-- Wait a few minutes
-- Cache prevents repeated requests (enabled by default)
 
 ---
 
-## 📚 File Structure
+ModuleNotFoundError
+
+pip install -r requirements.txt
 
-```
-stock-predictor/
-├── app.py                 # Gradio interface (main entry point)
-├── model.py              # LSTM model and prediction logic
-├── requirements.txt      # Python dependencies
-├── .env.example          # Environment variables template
-├── .gitignore            # Git ignore rules
-└── README.md             # This file
-
-Optional (generated):
-├── .env                  # Your API keys (NOT committed)
-├── lstm_model.pt         # Trained model cache
-├── prediction_cache.pkl  # Prediction cache
-└── __pycache__/          # Python cache
-```
 
 ---
 
-## 🤝 Contributing
+No Data Found
 
-Improvements welcome! Ideas:
-- Better sentiment analysis
-- Multiple models ensemble
-- Real-time streaming
-- Mobile app
-- Explainability (SHAP)
+Verify stock symbol format.
 
----
+Example:
 
-## 📄 License
+SBIN.NS
 
-MIT License - Free for personal and commercial use
 
 ---
 
-## 🔗 Resources
+📁 Project Structure
+
+stock-predictor/
+│
+├── app.py
+├── model.py
+├── requirements.txt
+├── README.md
+├── .env.example
+└── .gitignore
 
-- **Finnhub API**: https://finnhub.io/
-- **Gradio**: https://gradio.app/
-- **HF Spaces**: https://huggingface.co/spaces
-- **PyTorch**: https://pytorch.org/
-- **yfinance**: https://github.com/ranaroussi/yfinance
 
 ---
 
-## 📧 Support
+📄 License
+
+MIT License
 
-For issues:
-1. Check troubleshooting section above
-2. Verify API key is set
-3. Check stock symbol is valid
-4. Check internet connection
 
 ---
 
-**Made with ❤️ for stock enthusiasts and traders.**
+❤️ Built With
+
+Gradio
+
+PyTorch
+
+yfinance
+
+Finnhub API
 
-*Last updated: 2024*
+Python