用python，自编函数实现朴素贝叶斯分类算法。并对sklearn数据库中鸢尾花数据进行分类。输出分类结果，说明分类效果。

以下是一个简单的朴素贝叶斯分类算法的自编函数，并对sklearn数据库中的鸢尾花数据进行分类：

from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

def naive_bayes_classification(X_train, y_train, X_test):
    # 计算每个类别的先验概率
    classes = np.unique(y_train)
    prior_probs = {}
    for c in classes:
        prior_probs[c] = np.sum(y_train == c) / len(y_train)

    # 计算每个特征的条件概率
    feature_probs = {}
    for feature in range(X_train.shape[1]):
        feature_probs[feature] = {}
        unique_values = np.unique(X_train[:, feature])
        for value in unique_values:
            feature_probs[feature][value] = {}
            for c in classes:
                numerator = np.sum((X_train[:, feature] == value) & (y_train == c))
                denominator = np.sum(y_train == c)
                feature_probs[feature][value][c] = numerator / denominator

    # 进行分类
    y_pred = []
    for i in range(X_test.shape[0]):
        class_probs = {}
        for c in classes:
            class_prob = prior_probs[c]
            for feature in range(X_test.shape[1]):
                value = X_test[i, feature]
                if value in feature_probs[feature]:
                    class_prob *= feature_probs[feature][value][c]
                else:
                    class_prob *= 0.0001  # 平滑处理，避免概率为0
            class_probs[c] = class_prob
        pred_class = max(class_probs, key=class_probs.get)
        y_pred.append(pred_class)

    return y_pred

# 加载鸢尾花数据集
iris = load_iris()
X = iris.data
y = iris.target

# 划分训练集和测试集
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# 进行朴素贝叶斯分类
y_pred = naive_bayes_classification(X_train, y_train, X_test)

# 输出分类结果和分类效果
print("分类结果：", y_pred)
print("分类准确率：", accuracy_score(y_test, y_pred))

在这个示例中，我们首先使用`sklearn.datasets.load_iris`加载鸢尾花数据集。然后，我们使用`sklearn.model_selection.train_test_split`将数据集划分为训练集和测试集。接下来，我们调用自编函数`naive_bayes_classification`进行朴素贝叶斯分类。最后，使用`sklearn.metrics.accuracy_score`计算分类准确率，并输出分类结果和分类效果。

请注意，这只是一个简单的朴素贝叶斯分类算法实现，可能不是最优的方法。在实际应用中，可以考虑使用更复杂的模型或库来进行分类任务。