create_correlated_abilities

def create_correlated_abilities(correlation_matrix, n_participants):
    """ Creates correlated ability parameters based on an input correlation matrix.

    This is a helper function for use in synthesizing multi-dimensional data
    assuming multivariate normal distribution

    Args:
        correlation_matrix: (2d array) Symmetric matrix defining
                            the correlation between the abilities
        n_participants: number of participants to synthesize

    Returns:
        abilities: (2d array) correlated abilities
    """
    lower = np.linalg.cholesky(correlation_matrix)

    return lower @ np.random.randn(correlation_matrix.shape[0], n_participants)