This is a part of a code to achieve parameters for increasing the accuracy of prediction.
How can I get the parameters from the fitness funnction and COA function to put in the decision tree classifier?

def fitness_dt(parameters):
    # Unpack the parameters, ensuring they are appropriate integers
    max_depth = int(parameters[0])
    min_samples_split = int(parameters[1])
    min_samples_leaf = int(parameters[2])
    
    # Initialize and train the Decision Tree model using the provided parameters
    model = DecisionTreeClassifier(
        max_depth=max_depth if max_depth > 0 else None,  # None if max_depth is not positive
        min_samples_split=max(2, min_samples_split),    # At least 2
        min_samples_leaf=max(1, min_samples_leaf)       # At least 1
    )
    
    # Use cross-validation to evaluate the model
    scores = cross_val_score(model, X_train, y_train, cv=5, scoring='accuracy')
    
    # Return the negative of the accuracy as the fitness score to minimize
    return -np.mean(scores)



def coa_dt(searchagents, max_iterations, bounds, dimension, fitness_function):
    lowerbounds, upperbounds = np.array([b[0] for b in bounds]), np.array([b[1] for b in bounds])

    # INITIALIZATION
    x = np.random.rand(searchagents, dimension) * (upperbounds - lowerbounds) + lowerbounds
    x = np.round(x)  # Ensure parameters are integers
    fit = np.array([fitness_function(ind) for ind in x])

    # Variables to track the best solution found
    fbest = np.min(fit)
    xbest = x[np.argmin(fit), :]

    for t in range(max_iterations):
        for i in range(searchagents):
            # Exploration and Exploitation
            if np.random.rand() < 0.5:  # Explore
                iguana = xbest
                a = 1 + np.random.rand()
                x_p1 = x[i, :] + np.random.rand() * (iguana - a * x[i, :])
            else:  # Exploit
                lo_local = lowerbounds / (t+1)
                hi_local = upperbounds / (t+1)
                x_p1 = x[i, :] + (1 - 2 * np.random.rand()) * (lo_local + np.random.rand() * (hi_local - lo_local))

            # Ensure the new position is within bounds and is integer
            x_p1 = np.maximum(x_p1, lowerbounds)
            x_p1 = np.minimum(x_p1, upperbounds)
            x_p1 = np.round(x_p1)  # Ensure parameters are integers

            # Evaluate new position
            f_p1 = fitness_function(x_p1)
            if f_p1 < fit[i]:
                x[i, :] = x_p1
                fit[i] = f_p1

        # Update the best solution found
        current_fbest = np.min(fit)
        if current_fbest < fbest:
            fbest = current_fbest
            xbest = x[np.argmin(fit), :]

    return fbest, xbest

# Define bounds for Decision Tree hyperparameters
dt_bounds = [(1, 20), (2, 20), (1, 20)]  # Bounds for max_depth, min_samples_split, min_samples_leaf



opt_dt_model = DecisionTreeClassifier(max_depth=11, min_samples_split=9, min_samples_leaf=16, random_state=42)
opt_dt_model.fit(X_train, y_train)
y_pred_opt_dt = opt_dt_model.predict(X_test)
accuracy_opt_dt = accuracy_score(y_test, y_pred_opt_dt)
print("Accuracy:", accuracy_opt_dt)
print("Confusion Matrix:n", confusion_matrix(y_test, y_pred_opt_dt))

I dont know how can I find the parameters with these functions and put them in the classifier.
Please help me with that.