機器學習-分類模型

1 算法模型

我嘗試過的算法模型有：LogisticRegression-邏輯迴歸，KNeighborsClassifier-K近邻算法，GaussianProcessClassifier-高斯過程，RPF, GradientBoostingRegressor。算法具體內容可以百度搜索。

 

2 模型訓練

from sklearn.linear_model import LogisticRegression
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
from sklearn.externals import joblib
import os,cv2
import numpy as np
from sklearn.model_selection import cross_val_score
from collections import Counter
from sklearn.gaussian_process import GaussianProcessClassifier
from sklearn.gaussian_process.kernels import RBF
from sklearn.ensemble import GradientBoostingRegressor
import sklearn


def load_data(img_path):
    x = []
    y = []
    n = 0
    for file in os.listdir(img_path):
        file_path = os.path.join(img_path, file)
        for img in os.listdir(file_path):
            image_path = os.path.join(file_path, img)
            image = cv2.imread(image_path)
            input = image.reshape(1, -1)/255.
            x.append(input[0])
            y.append(n)
        n += 1
    x = np.array(x)
    y = np.array(y)
    return x, y

def train(x, y):
    print("starting train!")
    X_train,X_test,y_train,y_test=train_test_split(x,y,test_size=0.3)

    # Log = sklearn.svm.LinearSVC()
    # Log = sklearn.svm.SVC(kernel='linear',gamma='auto')
    # Log = sklearn.svm.NuSVC(gamma='auto')
    # Log = GradientBoostingRegressor(loss='quantile', alpha=0.95,
    #                                 n_estimators=250, max_depth=6,
    #                                 learning_rate=.1, min_samples_leaf=9,
    #                                 min_samples_split=9)
    # Log = GaussianProcessClassifier(1.0 * RBF(1.0))
    # Log = KNeighborsClassifier(3)
    # Log = LogisticRegression(penalty='l2',solver='lbfgs')
    Log = LogisticRegression(penalty='l2',dual=True,warm_start=True,)
    Log.fit(X_train,y_train)
    prediction = Log.predict(X_test)
    joblib.dump(Log,r'D:\MODEL\Y_IT_044_V1.model')
    print("Compete the training! The model is saved!")

    print("-"*100)
    print("starting eval!")
    classes = set(y_test)
    dict_ = {}
    for i in classes:
        dict_[i] = 0
    for i in range(len(y_test)):
        if y_test[i] != prediction[i]:
            dict_[y_test[i]] += 1
    for k in dict_.keys():
        acc = 1 - dict_[k]/Counter(y_test)[k]
        print('classes', k, 'eval acc: ', acc)


if __name__ == '__main__':
    img_path = r'D:\data'
    x, y = load_data(img_path)
    train(x, y)

3 模型測試

from sklearn.preprocessing import StandardScaler
import cv2, os
from sklearn.externals import joblib
import numpy as np


def model_Predict(img_path, model_path, classes):
    x = []
    for file in os.listdir(img_path):
        file_path = os.path.join(img_path, file)
        img = cv2.imread(file_path)
        # img = cv2.resize(img,(80,80))
        input = img.reshape(1, -1).astype(np.float64) / 255.
        x.append(input[0])
    x = np.array(x)
    Log = joblib.load(model_path)
    predict = Log.predict(x)
    count = 0
    for i in range(len(predict)):
        if predict[i] != classes:
            print(os.listdir(img_path)[i])
            count += 1
    acc = 1 - count / len(predict)
    print('acc:', acc)


if __name__ == '__main__':
    # 對某一個類別文件夾進行測試
    img_path = r'D:\data'
    model_path = r'D:\MODEL\R_IT_044_v0.model'
    classes = 1
    model_Predict(img_path, model_path, classes)