机器学习中根据训练组中及验证组的AUC均大于075且训练组AUC和验证组AUC差值小于01两个条件对SVM进行随机搜索调参若不满足条件则一直进行调参最终输出最佳参数模型python代码

以下是一种可能的Python代码实现：

from sklearn.svm import SVC
from sklearn.model_selection import RandomizedSearchCV
from sklearn.metrics import roc_auc_score
import numpy as np

# X_train, y_train, X_val, y_val为训练集和验证集的特征和标签
# param_distributions为SVM的参数分布范围
# n_iter为随机搜索的迭代次数
# scoring为评估指标，这里使用AUC作为评估指标
def random_search_SVM(X_train, y_train, X_val, y_val, param_distributions, n_iter=10, scoring='roc_auc'):
    best_score = 0
    while True:
        # 随机搜索SVM的参数
        svm = SVC()
        random_search = RandomizedSearchCV(svm, param_distributions, n_iter=n_iter, scoring=scoring)
        random_search.fit(X_train, y_train)
        
        # 计算训练集和验证集的AUC
        y_train_prob = random_search.predict_proba(X_train)[:, 1]
        train_auc = roc_auc_score(y_train, y_train_prob)
        y_val_prob = random_search.predict_proba(X_val)[:, 1]
        val_auc = roc_auc_score(y_val, y_val_prob)
        
        # 判断是否满足条件
        if train_auc > 0.75 and val_auc > 0.75 and abs(train_auc - val_auc) < 0.1:
            # 输出最佳参数模型
            print('Best parameters:', random_search.best_params_)
            print('Best score:', random_search.best_score_)
            return random_search.best_estimator_
        
        # 记录当前最佳结果
        if val_auc > best_score:
            best_score = val_auc
            best_estimator = random_search.best_estimator_
``