Hi Daniel, I've experimented with some ML models before, and I found it quite challenging to make accurate predictions within a short timeframe, like 1 minute. These types of models are typically more effective in high-frequency trading scenarios. If you're interested in learning more, I've attached a PDF that provides additional details.

good work @Amit Ranjan

import numpy as np import pandas as pd import matplotlib.pyplot as plt def calculate_gbm(S0, T, mu, sigma, n_steps, n_sims): dt = T / n_steps gbm_paths = np.zeros((n_steps + 1, n_sims)) gbm_paths[0] = S0 for t in range(1, n_steps + 1): Z = np.random.standard_normal(n_sims) gbm_paths[t] = gbm_paths[t - 1] * np.exp((mu - 0.5 * sigma**2) * dt + sigma * np.sqrt(dt) * Z) return gbm_paths def main(): closing_prices=pd.read_csv(r"near_data.csv", parse_dates=['DATE'])['PRICE'].dropna() print(closing_prices) # Calculate log returns log_returns = np.log(closing_prices / closing_prices.shift(1)).dropna() mu = log_returns.mean() sigma = log_returns.std() # Simulation parameters S0 = closing_prices.iloc[-1] T = 1/12 # 1 year n_steps = 30 # Number of trading days in a year n_sims = 365 # Number of simulations gbm_paths = calculate_gbm(S0, T, mu, sigma, n_steps, n_sims) # Plotting the simulations plt.figure(figsize=(10, 6)) plt.plot(gbm_paths) plt.title(f'Geometric Brownian Motion ') plt.xlabel('Days') plt.ylabel('Stock Price') plt.show() if _name_ == '_main_': main()