python 模型保存问题

"""
   随机森林模型
"""


from sklearn.model_selection import train_test_split, GridSearchCV

from sklearn.feature_extraction import DictVectorizer

from sklearn.ensemble import RandomForestClassifier
import pandas as pd
import numpy as np
import os
import joblib
import time
from sklearn.preprocessing import MinMaxScaler








def decision():
    """
    随机森林对电话号码做分类
    :return: None
    """
    # 获取数据
    all_data = pd.read_csv('e:/index/t.ns300.csv')

    # 处理数据，找出特征值和目标值
    x = all_data[[ 'CALLTIMES',        'CALLEDNUMBERS',    'CALLEDLSD',    'PJTALKLEN',    'TOTALCALLLEN',    'BZCTALKLEN',    'NOTALKTIMES',    'JTL',    'CALLING_PROVINCE_MODE',    'CALLING_CITY_MODE',    'CALLED_PROVINCE_MODE',    'CALLED_CITY_MODE',    'CALLEN_PJ',    'YIDI_PROVINCE_PER',    'YIDI_CITY_PER']]

    y = all_data['TOPIC_CLASS']

    print(x)
    # 异常值处理
    # x['JTL'].fillna(x['JTL'].mean(), inplace=True)

    # 分割数据集到训练集合测试集
    x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.25)

    # 进行处理（特征工程）
    dict = DictVectorizer(sparse=False)

    x_train = dict.fit_transform(x_train.to_dict(orient="records"))

    print(dict.get_feature_names())

    x_test = dict.transform(x_test.to_dict(orient="records"))

    # 随机森林进行预测 （超参数调优）
    rf = RandomForestClassifier()

    param = {"n_estimators": [50,100,120, 200, 300, 500, 800, 1200,1500], "max_depth": [5, 8,10, 15,20, 25, 30,35]}

    # 网格搜索与交叉验证
    gc = GridSearchCV(rf, param_grid=param, cv=5)

    gc.fit(x_train, y_train)

    print("准确率：", gc.score(x_test, y_test))

    if os.path.exists("e:/index/result/test_ns300_data.csv"):

        os.remove("e:/index/result/test_ns300_data.csv")

    np.savetxt("e:/index/result/test_ns300_data.csv",x_test,fmt='%3f',delimiter=',')
    if os.path.exists("e:/index/result/realty_ns300_data.csv"):

        os.remove("e:/index/result/realty_ns300_data.csv")

    np.savetxt("e:/index/result/realty_ns300_data.csv",y_test,header='realty')

    print("查看选择的参数模型：", gc.best_params_)

    print("预测结果：",gc.predict(x_test))

    if os.path.exists("e:/index/result/predict_ns300_data.csv"):

        os.remove("e:/index/result/predict_ns300_data.csv")

    np.savetxt("e:/index/result/predict_ns300_data.csv",gc.predict(x_test),header='predict')

    # 创建文件目录
    dirs = 'e:/index/testModel'

    if not os.path.exists(dirs):

        os.makedirs(dirs)

    # 保存模型
    joblib.dump(gc, dirs + '/rf_ns300.pkl')

    return None





if __name__ == "__main__":
    start=time.perf_counter()
    decision()
    end=time.perf_counter()
    print('模块运行时间: %s Seconds'%(end-start))