self-train RandomForesrClassifier

随便玩一下啊

---

感受野绘制

写论文的需要就随便写了个小玩意，后面再继续更新

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.ensemble import RandomForestRegressor
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split
import time
file = r'D:\微信\WeChat Files\wxid_8f01i8xhk84722\FileStorage\File\2021-09\identity_0829.csv'
data = pd.read_csv(file)
# data = np.random.shuffle(data)
data = np.array(data)
data_x = data[:,6:9]
data_y = data[:,9]-1

#测试集的比例占80%
x_train,x_val,y_train,y_val = train_test_split(data_x,data_y,test_size=0.99)
#训练集当中有标签的数据占20%，无标签占80%
x_train_labeled,x_train_unlabeled,y_train_labeled,y_train_unlabeled = train_test_split(x_train,y_train,test_size=0.9)

rf_cls = RandomForestClassifier()

def train(x_train,y_train,x_unlabeled,cls,x_val,y_val):
    plt.ion()
    for i in range(20):
        print('第%d次迭代'%(i+1))
        print('训练数据：', len(x_train))
        print('无标签数据：', len(x_unlabeled))
        cls.fit(x_train, y_train.astype('int'))
        print('模型得分：',cls.score(x_val,y_val.astype('int')))
        score = cls.predict_proba(x_unlabeled)
        label = cls.predict(x_unlabeled)
        score = np.max(score, axis=1)
        index = score > 0.999
        print('新加入数据：', len(x_unlabeled[index]),'\n')
        # print(x_train.shape,x_unlabeled[index].shape)
        x_train = np.concatenate((x_train, x_unlabeled[index]),axis=0)
        y_train = np.concatenate((y_train, label[index]),axis=0)
        x_unlabeled = x_unlabeled[~index]
        # print(x_train.shape,y_train.shape)
        plot(x_train[:,0],x_train[:,1],x_train[:,2],y_train)
        # plt.scatter(x_train[:,0], x_train[:,1], c=y_train)
        # plt.pause(0.5)

    plt.ioff()
    plt.show()

def plot(x,y,z,label):
    # x,y = np.meshgrid(x,y)
    ax = plt.axes(projection='3d')
    ax.scatter(x,y,z,c=label, cmap='viridis',label=label)
    plt.pause(1)
    plt.legend(label)
    # plt.show()

train(x_train_labeled,y_train_labeled,x_train_unlabeled,rf_cls,x_val,y_val)

总结

老老实实的迈好每一步，相信成功就在你身边 甘愿为理想“头破血流”