Update README.md

README.md

@@ -158,4 +158,178 @@ configs:
   data_files:
   - split: train
     path: data/train-*
+license: mit
+task_categories:
+- time-series-forecasting
+- reinforcement-learning
+- tabular-regression
+language:
+- en
+tags:
+- finance
+- time-series
+- stocks
+- technical-analysis
+- yahoo-finance
+- reinforcement-learning
+pretty_name: S&P 500 Comprehensive Stock Market Dataset
+size_categories:
+- 100K<n<1M
 ---
+S&P 500 Comprehensive Stock Market Dataset
+Dataset Description
+This dataset contains comprehensive historical stock market data for S&P 500 companies, spanning the last 5 years with extensive feature engineering and technical analysis indicators. The dataset is designed for time series forecasting, stock price prediction, and financial modeling tasks.
+
+Dataset Summary
+Total Records: 620,095 daily observations
+Features: 73 comprehensive features including raw price data, technical indicators, and engineered features
+Time Period: Last 5 years of historical data
+Data Source: Yahoo Finance API
+Companies: S&P 500 constituent stocks
+Update Frequency: Daily market data
+Features Overview
+Basic Market Data (9 features)
+Date: Trading date timestamp
+Open, High, Low, Close: Standard OHLC price data
+Volume: Number of shares traded
+Dividends: Dividend payments on the date
+Stock Splits: Stock split information
+Ticker: Stock symbol identifier
+Technical Analysis Indicators (16 features)
+Moving Averages
+SMA_5, SMA_10, SMA_20, SMA_50: Simple Moving Averages for different periods
+EMA_12, EMA_26: Exponential Moving Averages
+Momentum Indicators
+MACD, MACD_Signal, MACD_Histogram: Moving Average Convergence Divergence indicators
+RSI: Relative Strength Index (14-period)
+Volatility Indicators
+BB_Middle, BB_Upper, BB_Lower: Bollinger Bands components
+BB_Width: Bollinger Bands width
+BB_Position: Price position within Bollinger Bands
+Volatility: Historical volatility measure
+Engineered Features (16 features)
+Price_Change: Daily price change
+Price_Change_5d: 5-day price change
+High_Low_Ratio: Ratio of high to low prices
+Open_Close_Ratio: Ratio of open to close prices
+Volume_SMA: Volume simple moving average
+Volume_Ratio: Current volume to average volume ratio
+Lagged Features (32 features)
+The dataset includes lagged versions of key features for time series modeling:
+
+Price Lags (5 features)
+Close_lag_1 to Close_lag_10: Historical closing prices
+Volume Lags (5 features)
+Volume_lag_1 to Volume_lag_10: Historical volume data
+Price Change Lags (5 features)
+Price_Change_lag_1 to Price_Change_lag_10: Historical price changes
+RSI Lags (5 features)
+RSI_lag_1 to RSI_lag_10: Historical RSI values
+MACD Lags (5 features)
+MACD_lag_1 to MACD_lag_10: Historical MACD values
+Volatility Lags (5 features)
+Volatility_lag_1 to Volatility_lag_10: Historical volatility measures
+Target Variables (12 features)
+The dataset includes multiple prediction targets for different time horizons:
+
+1-Day Predictions
+Future_Return_1d: Next day return percentage
+Future_Up_1d: Binary indicator for price increase
+Future_Category_1d: Categorical classification of price movement
+5-Day Predictions
+Future_Return_5d: 5-day forward return
+Future_Up_5d: 5-day binary direction
+Future_Category_5d: 5-day categorical movement
+10-Day Predictions
+Future_Return_10d: 10-day forward return
+Future_Up_10d: 10-day binary direction
+Future_Category_10d: 10-day categorical movement
+20-Day Predictions
+Future_Return_20d: 20-day forward return
+Future_Up_20d: 20-day binary direction
+Future_Category_20d: 20-day categorical movement
+Use Cases
+Primary Applications
+Stock Price Prediction: Forecast future stock prices using technical indicators and historical patterns
+Direction Classification: Predict whether stock prices will go up or down
+Risk Assessment: Analyze volatility patterns and market risk
+Trading Strategy Development: Backtest and develop algorithmic trading strategies
+Financial Research: Academic research in computational finance and market efficiency
+Machine Learning Tasks
+Regression: Predict continuous returns (Future_Return_*)
+Binary Classification: Predict price direction (Future_Up_*)
+Multi-class Classification: Predict categorical movements (Future_Category_*)
+Time Series Forecasting: Leverage lagged features for temporal modeling
+Anomaly Detection: Identify unusual market patterns
+Data Quality and Processing
+Feature Engineering Quality
+Technical Indicators: Industry-standard calculations for all technical analysis features
+Lag Features: Comprehensive historical context with multiple time horizons
+Target Engineering: Multiple prediction horizons for flexible modeling approaches
+Data Validation: Processed through robust data validation pipelines
+Missing Data Handling
+Forward-fill methodology for missing price data
+Proper handling of weekends and market holidays
+Corporate action adjustments included
+Data Integrity
+All features computed using vectorized operations for consistency
+Proper handling of stock splits and dividend adjustments
+No look-ahead bias in feature construction
+Usage Examples
+Loading the Dataset
+python
+from datasets import load_dataset
+
+dataset = load_dataset("Adilbai/stock-dataset")
+df = dataset["train"].to_pandas()
+Basic Analysis
+python
+# View dataset structure
+print(f"Dataset shape: {df.shape}")
+print(f"Date range: {df['Date'].min()} to {df['Date'].max()}")
+print(f"Number of unique tickers: {df['Ticker'].nunique()}")
+
+# Check target distributions
+print(df['Future_Up_1d'].value_counts())
+Feature Selection for Modeling
+python
+# Technical indicators
+technical_features = ['SMA_5', 'SMA_10', 'RSI', 'MACD', 'BB_Position', 'Volatility']
+
+# Lagged features
+lag_features = [col for col in df.columns if 'lag' in col]
+
+# Targets
+targets = ['Future_Return_1d', 'Future_Up_1d', 'Future_Category_1d']
+Dataset Statistics
+Coverage: All S&P 500 constituent companies
+Temporal Span: 5 years of daily data
+Feature Density: 73 features per observation
+Target Variety: 12 different prediction targets
+Technical Indicators: 16 professional-grade technical analysis features
+Lag Depth: Up to 10-period historical context
+Limitations and Considerations
+Data Limitations
+Survivorship Bias: Only includes current S&P 500 constituents
+Market Hours: Only includes regular trading session data
+Corporate Actions: Historical adjustments may affect long-term patterns
+Usage Considerations
+Look-ahead Bias: Ensure proper train/test splits respecting temporal order
+Market Regime Changes: Model performance may vary across different market conditions
+Feature Correlation: Many technical indicators are derived from the same underlying price data
+Citation
+If you use this dataset in your research, please cite:
+
+bibtex
+@dataset{adilbai_sp500_dataset,
+  title={S&P 500 Comprehensive Stock Market Dataset},
+  author={Adilbai},
+  year={2024},
+  publisher={Hugging Face},
+  url={https://huggingface.co/datasets/Adilbai/stock-dataset}
+}
+License
+This dataset is released under the MIT License. Please note that while the dataset compilation and feature engineering are provided under MIT license, users should be aware of Yahoo Finance's terms of service for the underlying data.
+
+Disclaimer
+This dataset is provided for educational and research purposes only. It should not be used as the sole basis for investment decisions. Past performance does not guarantee future results. Users should conduct their own research and consider consulting with financial advisors before making investment decisions.