吴恩达Coursera, 机器学习专项课程, Machine Learning：Unsupervised Learning, Recommenders, Reinforcement Learning第一周编程作业2

吴恩达Coursera, 机器学习专项课程, Machine Learning：Unsupervised Learning, Recommenders, Reinforcement Learning第一周所有jupyter notebook文件2：

吴恩达Coursera, 机器学习专项课程, Machine Learning：Unsupervised Learning, Recommenders, Reinforcement Learning第一周所有jupyter notebook文件(包括实验室练习文件)1

本次作业

Exercise 1

# UNQ_C1
# GRADED FUNCTION: estimate_gaussian
 
def estimate_gaussian(X): 
    """
    Calculates mean and variance of all features 
    in the dataset
    
    Args:
        X (ndarray): (m, n) Data matrix
    
    Returns:
        mu (ndarray): (n,) Mean of all features
        var (ndarray): (n,) Variance of all features
    """
 
    m, n = X.shape
    
    ### START CODE HERE ### 
    mu = np.zeros((n,1))
    var = np.zeros((n,1))
 
    mu = np.mean(X, axis=0) # axis=0表示列，每列的均值
    var = np.var(X,axis=0) # 求每列的方差    
    
    ### END CODE HERE ### 
        
    return mu, var

Exercise 2

# UNQ_C2
# GRADED FUNCTION: select_threshold
 
def select_threshold(y_val, p_val): 
    """
    Finds the best threshold to use for selecting outliers 
    based on the results from a validation set (p_val) 
    and the ground truth (y_val)
    
    Args:
        y_val (ndarray): Ground truth on validation set
        p_val (ndarray): Results on validation set
        
    Returns:
        epsilon (float): Threshold chosen 
        F1 (float):      F1 score by choosing epsilon as threshold
    """ 
 
    best_epsilon = 0
    best_F1 = 0
    F1 = 0
    
    step_size = (max(p_val) - min(p_val)) / 1000
    
    for epsilon in np.arange(min(p_val), max(p_val), step_size):
    
        ### START CODE HERE ### 
        cvPrecision = p_val < epsilon
        tp = np.sum((cvPrecision == 1) & (y_val == 1)).astype(float)  # sum求和是int型的，需要转为float
        fp = np.sum((cvPrecision == 1) & (y_val == 0)).astype(float)
        fn = np.sum((cvPrecision == 0) & (y_val == 1)).astype(float)
        precision = tp/(tp+fp)  # 精准度
        recision = tp/(tp+fn)   # 召回率
        F1 = (2*precision*recision)/(precision+recision)  # F1Score计算公式            
        ### END CODE HERE ###    
        if F1 > best_F1:
            best_F1 = F1
            best_epsilon = epsilon
        
    return best_epsilon, best_F1