example.config.yml

# Multi-Modal Sentiment Analysis for Crypto Market Prediction Configuration

# Cryptocurrency symbols to analyze
symbols:
  - BTC/USD
  - ETH/USD
  - XRP/USD

# Market data settings
market_data:
  timeframe: 1h  # 1 hour candles
  lookback_period: 168  # 7 days of hourly data

# Sentiment analysis settings
sentiment:
  window: 24h  # Time window for sentiment analysis
  update_interval: 15m  # How often to update sentiment scores

# Twitter API settings
twitter:
  max_tweets: 1000  # Maximum number of tweets to collect per symbol
  languages: 
    - en
    - es

# Reddit API settings
reddit:
  subreddits:
    - CryptoCurrency
    - Bitcoin
    - ethereum
  post_limit: 500  # Number of posts to analyze per subreddit

# Model parameters
model_params:
  market_input_dim: 168  # Should match lookback_period
  sentiment_input_dim: 24  # Should match sentiment window in hours
  lstm_units: 64
  dense_units: [128, 64]
  dropout_rate: 0.2

# Training settings
training:
  batch_size: 32
  epochs: 100
  validation_split: 0.2
  early_stopping_patience: 10

# Prediction settings
prediction:
  interval: 1h  # How often to make new predictions

# Logging settings
logging:
  level: INFO
  file: logs/mmsag.log

# TensorBoard settings
tensorboard:
  log_dir: logs/tensorboard

# Dashboard settings
dashboard:
  update_interval: 5m  # How often to update the dashboard
  prediction_days: 7  # Number of days to predict in the future