import pandas as pd import yfinance as yf import matplotlib.pyplot as plt import numpy as np # Define Nadaraya-Watson Envelope function def nadaraya_watson_envelope(price_data, bandwidth=0.1): smoothed_values = [] for i in range(len(price_data)): xi = price_data.iloc[i]['Close'] weights = [] for j in range(len(price_data)): xj = price_data.iloc[j]['Close'] kernel_weight = np.exp(-(xi - xj) ** 2 / (2 * bandwidth ** 2)) weights.append(kernel_weight) weighted_prices = price_data['Close'] * weights smoothed_value = weighted_prices.sum() / sum(weights) smoothed_values.append(smoothed_value) upper_envelope = np.array(smoothed_values) + bandwidth lower_envelope = np.array(smoothed_values) - bandwidth return upper_envelope, lower_envelope # Generate signals def generate_signals(price_data, bandwidth=0.1): upper_envelope, lower_envelope = nadaraya_watson_envelope(price_data, bandwidth) signals = [] for i in range(len(price_data)): if price_data['Close'][i] > upper_envelope[i]: signals.append('SELL') elif price_data['Close'][i] < lower_envelope[i]: signals.append('BUY') else: signals.append('HOLD') return signals # Load historical data data = yf.download('GOOGL', start="2020-01-01", end="2022-04-30") # Extract the closing prices close_prices = data['Close'] # Set the bandwidth for Nadaraya-Watson Envelopes bandwidth = 10.0 # Adjust this value as needed # Calculate the Nadaraya-Watson Envelopes upper_envelope, lower_envelope = nadaraya_watson_envelope(close_prices, bandwidth) # Generate trading signals signals = generate_signals(close_prices, bandwidth) # Implement risk management and portfolio simulation def simulate_portfolio(signals, initial_cash=100000): portfolio = pd.DataFrame(index=signals.index) portfolio['Signal'] = signals['BuySignal'] + signals['SellSignal'] portfolio['Position'] = portfolio['Signal'].cumsum() portfolio['Cash'] = initial_cash - (portfolio['Signal'] * signals['Price']).cumsum() # Calculate portfolio value portfolio['PortfolioValue'] = portfolio['Cash'] + (portfolio['Position'] * signals['Price']) return portfolio # Simulate the portfolio portfolio = simulate_portfolio(signals) # Plot the results plt.figure(figsize=(12, 8)) plt.title('Nadaraya-Watson Envelopes Trading Strategy') plt.plot(signals.index, close_prices, label='Close Price', lw=1) plt.plot(signals.index, upper_envelope, 'g', label='Upper Envelope', lw=1) plt.plot(signals.index, lower_envelope, 'r', label='Lower Envelope', lw=1) plt.legend() plt.figure(figsize=(12, 4)) plt.title('Portfolio Performance') plt.plot(portfolio.index, portfolio['PortfolioValue'], label='Portfolio Value', lw=1) plt.legend() plt.show()