Upload LSTM stock prediction model with sentiment analysis

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+stage2_universal_lstm_20250705_170829.keras filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,97 @@
+---
+library_name: tensorflow
+tags:
+- lstm
+- stock-prediction
+- time-series
+- financial-analysis
+- tensorflow
+- keras
+pipeline_tag: tabular-regression
+---
+
+# Universal LSTM Stock Prediction Model
+
+## Model Description
+
+This is a Universal LSTM model trained for stock price prediction with sentiment analysis integration. The model can predict stock prices for multiple stocks using technical indicators and sentiment features.
+
+## Model Details
+
+- **Model Type**: LSTM (Long Short-Term Memory) Neural Network
+- **Framework**: TensorFlow/Keras
+- **Task**: Time Series Prediction / Stock Price Forecasting
+- **Training Date**: July 5, 2025
+- **Input Features**: Technical indicators + Sentiment scores
+- **Output**: Next-day stock price prediction
+
+## Architecture
+
+- **Input Shape**: (None, 60, 6)
+- **Output Shape**: (None, 1)
+- **Parameters**: 31,651
+
+The model uses a multi-layer LSTM architecture optimized for financial time series prediction.
+
+## Training Data
+
+- **Stocks**: Multiple stocks including AAPL, MSFT, GOOGL, AMZN, TSLA, META, NVDA
+- **Features**: 
+  - Technical indicators (Open, High, Low, Close, Volume)
+  - Sentiment scores from financial news analysis
+  - Moving averages and other derived features
+- **Time Period**: Historical stock data with corresponding news sentiment
+- **Preprocessing**: StandardScaler normalization
+
+## Performance
+
+The model has been evaluated on multiple stocks with varying performance metrics. Best performance achieved:
+- **RMSE**: Varies by stock (typically 5-15 points)
+- **Architecture**: Optimized through hyperparameter tuning
+
+## Usage
+
+```python
+import tensorflow as tf
+import pickle
+import numpy as np
+
+# Load the model
+model = tf.keras.models.load_model('stage2_universal_lstm_20250705_170829.keras')
+
+# Load the scaler
+with open('stage2_scalers_20250705_170829.pkl', 'rb') as f:
+    scalers = pickle.load(f)
+
+# Prepare your data (X should be scaled using the same scaler)
+# X = your_data  # Shape: (batch_size, sequence_length, features)
+# predictions = model.predict(X)
+```
+
+## Files Included
+
+- `stage2_universal_lstm_20250705_170829.keras`: Main model file
+- `stage2_scalers_20250705_170829.pkl`: Feature scalers for preprocessing
+- `stage2_metadata_20250705_170829.json`: Model metadata and configuration
+- `stage2_architecture_20250705_170829.txt`: Detailed architecture description
+
+## Citation
+
+If you use this model in your research, please cite:
+
+```
+@misc{lstm_stock_prediction_2025,
+  title={Universal LSTM Stock Prediction Model with Sentiment Analysis},
+  author={Jeng Yang},
+  year={2025},
+  url={https://huggingface.co/jengyang/lstm-stock-prediction-model}
+}
+```
+
+## Disclaimer
+
+This model is for research and educational purposes only. Stock market predictions are inherently uncertain and this model should not be used as the sole basis for financial decisions. Always consult with financial professionals and conduct your own research before making investment decisions.
+
+## Contact
+
+For questions or issues, please open an issue in the model repository.

stage2_architecture_20250705_170829.txt

@@ -0,0 +1,65 @@
+
+Stage 2 Simple Universal LSTM Model Architecture
+=============================================
+
+Model Type: Sequential LSTM
+Training Date: 20250705_170829
+Training Stocks: 20 stocks
+Features: 6 features
+Target: Close
+
+Architecture:
+1. LSTM Layer 1: 50 units, return_sequences=True
+2. Dropout: 0.2
+3. LSTM Layer 2: 50 units, return_sequences=False  
+4. Dropout: 0.2
+5. Dense Output: 1 unit
+
+Training Parameters:
+- Batch Size: 16
+- Learning Rate: 0.001
+- Epochs Trained: 88
+- Early Stopping Patience: 12
+
+Performance Metrics:
+- RMSE: $26.85
+- MAE: $12.60
+- R�: 0.9903
+- MAPE: 2.68%
+
+Features Used:
+- Open
+- High
+- Low
+- Close
+- Volume
+- sentiment_10d_avg
+
+Stocks Trained On:
+- AAPL
+- AMZN
+- AVGO
+- BRK.B
+- COST
+- GOOG
+- JNJ
+- JPM
+- LLY
+- MA
+- META
+- MSFT
+- NFLX
+- NVDA
+- ORCL
+- PG
+- TSLA
+- V
+- WMT
+- XOM
+
+Usage Instructions:
+1. Load model: model = tf.keras.models.load_model('../models\stage2_universal_lstm_20250705_170829.keras')
+2. Load scalers: scalers = joblib.load('../models\stage2_scalers_20250705_170829.pkl')
+3. Preprocess data using feature_scaler and target_scaler
+4. Make predictions on sequences of shape (batch_size, 60, 6)
+5. Inverse transform predictions using target_scaler

stage2_metadata_20250705_170829.json

@@ -0,0 +1,49 @@
+{
+  "model_name": "Stage 2 Simple Universal LSTM",
+  "stocks": [
+    "AAPL",
+    "AMZN",
+    "AVGO",
+    "BRK.B",
+    "COST",
+    "GOOG",
+    "JNJ",
+    "JPM",
+    "LLY",
+    "MA",
+    "META",
+    "MSFT",
+    "NFLX",
+    "NVDA",
+    "ORCL",
+    "PG",
+    "TSLA",
+    "V",
+    "WMT",
+    "XOM"
+  ],
+  "features": [
+    "Open",
+    "High",
+    "Low",
+    "Close",
+    "Volume",
+    "sentiment_10d_avg"
+  ],
+  "target": "Close",
+  "time_steps": 60,
+  "batch_size": 16,
+  "epochs_trained": 88,
+  "final_val_loss": 0.0008096768870018423,
+  "rmse": 26.850322364411017,
+  "mae": 12.60320354323313,
+  "r2": 0.9902821175992801,
+  "mape": 2.679250893651326,
+  "created_at": "20250705_170829",
+  "architecture": {
+    "lstm1_units": 50,
+    "lstm2_units": 50,
+    "dropout_rate": 0.2,
+    "dense_units": 1
+  }
+}

stage2_scalers_20250705_170829.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:854aed206979a676a362470fb1137e86e16cfa51a46b10f7f918f1cba35cfa3b
+size 1289

stage2_universal_lstm_20250705_170829.keras

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e5844c9c50e36dc3df1495111e66b06babaa061400125bb6371939b34735192
+size 416407