\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6：\u901a\u8fc7AI\u589e\u5f3a\u6d4b\u8bd5\u811a\u672c\u751f\u6210\u4e0e\u7ef4\u62a4<\/p>\n<\/li>\n
\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b：\u5e94\u7528\u673a\u5668\u5b66\u4e60\u81ea\u52a8\u8bc6\u522b\u8f6f\u4ef6\u7f3a\u9677<\/p>\n<\/li>\n
A\/B\u6d4b\u8bd5\u4f18\u5316：\u4f7f\u7528AI\u7b97\u6cd5\u52a0\u901f\u5b9e\u9a8c\u51b3\u7b56\u8fc7\u7a0b<\/p>\n<\/li>\n<\/ul>\n
\u901a\u8fc7\u6280\u672f\u5206\u6790\u3001\u4ee3\u7801\u5b9e\u73b0\u548c\u53ef\u89c6\u5316\u5c55\u793a，\u6211\u4eec\u5c06\u63ed\u793aAI\u5982\u4f55\u63d0\u5347\u6d4b\u8bd5\u6548\u738750%\u4ee5\u4e0a，\u51cf\u5c11\u7f3a\u9677\u9003\u9038\u738740%，\u5e76\u4f18\u5316A\/B\u6d4b\u8bd5\u8d44\u6e90\u5206\u914d\u3002<\/p>\n
2. \u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u7684AI\u8fdb\u5316<\/h3>\n2.1 \u4f20\u7edf\u6846\u67b6\u7684\u5c40\u9650\u6027<\/h4>\n
graph LR
\n\u00a0 \u00a0 A[\u624b\u5de5\u6d4b\u8bd5] –> B[\u5f55\u5236\u56de\u653e\u5de5\u5177]
\n\u00a0 \u00a0 B –> C[\u811a\u672c\u5316\u6846\u67b6]
\n\u00a0 \u00a0 C –> D[\u6570\u636e\u9a71\u52a8\u6846\u67b6]
\n\u00a0 \u00a0 D –> E[\u5173\u952e\u5b57\u9a71\u52a8\u6846\u67b6]
\n\u00a0 \u00a0 E –> F[AI\u589e\u5f3a\u6846\u67b6]<\/p>\n
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/07\/20250730120223-688a09cfbb7e2.png\")
<\/p>\n
2.2 AI\u589e\u5f3a\u7684\u5173\u952e\u6280\u672f<\/h4>\n\n- \n
\u667a\u80fd\u5143\u7d20\u5b9a\u4f4d：\u81ea\u9002\u5e94UI\u53d8\u5316<\/p>\n<\/li>\n
- \n
\u6d4b\u8bd5\u7528\u4f8b\u751f\u6210：\u57fa\u4e8e\u4ee3\u7801\u5206\u6790\u7684\u81ea\u52a8\u6d4b\u8bd5\u751f\u6210<\/p>\n<\/li>\n
- \n
\u81ea\u6108\u673a\u5236：\u81ea\u52a8\u4fee\u590d\u5931\u6548\u6d4b\u8bd5<\/p>\n<\/li>\n
- \n
\u89c6\u89c9\u9a8c\u8bc1：\u57fa\u4e8eCV\u7684UI\u6bd4\u8f83<\/p>\n<\/li>\n<\/ul>\n
2.3 Python\u5b9e\u73b0：AI\u9a71\u52a8\u7684Selenium\u589e\u5f3a\u6846\u67b6<\/h4>\n
python<\/p>\n
from selenium import webdriver
\nfrom selenium.webdriver.common.by import By
\nfrom selenium.common.exceptions import NoSuchElementException
\nfrom sklearn.ensemble import RandomForestClassifier
\nimport cv2
\nimport numpy as np
\nimport time<\/p>\n
class AITestDriver:
\n def __init__(self):
\n self.driver = webdriver.Chrome()
\n self.element_classifier = self.load_element_classifier()
\n self.screenshot_history = []<\/p>\n
def load_element_classifier(self):
\n # \u5b9e\u9645\u5e94\u7528\u4e2d\u5e94\u4ece\u6301\u4e45\u5316\u5b58\u50a8\u52a0\u8f7d\u9884\u8bad\u7ec3\u6a21\u578b
\n return RandomForestClassifier()<\/p>\n
def smart_find_element(self, element_type, context):
\n """\u4f7f\u7528\u673a\u5668\u5b66\u4e60\u6a21\u578b\u5b9a\u4f4d\u5143\u7d20"""
\n try:
\n # \u4f20\u7edf\u5b9a\u4f4d\u65b9\u5f0f
\n return self.driver.find_element(By.XPATH, context)
\n except NoSuchElementException:
\n # \u4f7f\u7528CV\u548cML\u8fdb\u884c\u667a\u80fd\u5b9a\u4f4d
\n screenshot = self.capture_screenshot()
\n return self.ai_element_detection(screenshot, element_type)<\/p>\n
def ai_element_detection(self, image, element_type):
\n """\u57fa\u4e8e\u8ba1\u7b97\u673a\u89c6\u89c9\u7684\u5143\u7d20\u68c0\u6d4b"""
\n # \u7b80\u5316\u7684\u5143\u7d20\u68c0\u6d4b\u903b\u8f91
\n gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
\n edges = cv2.Canny(gray, 50, 150)<\/p>\n
# \u5b9e\u9645\u5e94\u7528\u4e2d\u5e94\u4f7f\u7528\u8bad\u7ec3\u597d\u7684\u6a21\u578b\u8fdb\u884c\u9884\u6d4b
\n # element_position = self.element_classifier.predict(edges)<\/p>\n
# \u793a\u4f8b\u8fd4\u56de
\n return MockElement()<\/p>\n
def visual_validation(self, current_screen):
\n """\u57fa\u4e8e\u89c6\u89c9\u7684UI\u9a8c\u8bc1"""
\n if not self.screenshot_history:
\n self.screenshot_history.append(current_screen)
\n return True<\/p>\n
# \u8ba1\u7b97\u4e0e\u5386\u53f2\u622a\u56fe\u7684\u5dee\u5f02
\n prev = cv2.cvtColor(self.screenshot_history[-1], cv2.COLOR_BGR2GRAY)
\n curr = cv2.cvtColor(current_screen, cv2.COLOR_BGR2GRAY)
\n diff = cv2.absdiff(prev, curr)<\/p>\n
# \u5982\u679c\u5dee\u5f02\u8d85\u8fc7\u9608\u503c\u5219\u6807\u8bb0\u4e3a\u5931\u8d25
\n if np.mean(diff) > 5.0:
\n cv2.imwrite('visual_diff.png', diff)
\n return False
\n return True<\/p>\n
def capture_screenshot(self):
\n return np.array(self.driver.get_screenshot_as_png())<\/p>\n
class MockElement:
\n def click(self):
\n print("AI\u5b9a\u4f4d\u5143\u7d20\u88ab\u70b9\u51fb")<\/p>\n
# \u4f7f\u7528\u793a\u4f8b
\ndriver = AITestDriver()
\ndriver.driver.get("https:\/\/example.com")
\nlogin_button = driver.smart_find_element("button", "\/\/button[@id='login']")
\nlogin_button.click()<\/p>\n
current_screen = driver.capture_screenshot()
\nif not driver.visual_validation(current_screen):
\n print("UI\u89c6\u89c9\u9a8c\u8bc1\u5931\u8d25！")<\/p>\n
2.4 \u6846\u67b6\u6027\u80fd\u5bf9\u6bd4<\/h4>\n
\n\u6307\u6807\u4f20\u7edf\u6846\u67b6AI\u589e\u5f3a\u6846\u67b6\u6539\u8fdb\u5e45\u5ea6<\/tr>\n \n\n\u6d4b\u8bd5\u521b\u5efa\u65f6\u95f4<\/td>\n 8\u5c0f\u65f6<\/td>\n 2\u5c0f\u65f6<\/td>\n -75%<\/td>\n<\/tr>\n \n\u7ef4\u62a4\u6210\u672c<\/td>\n \u9ad8<\/td>\n \u4f4e<\/td>\n -60%<\/td>\n<\/tr>\n \n\u5143\u7d20\u5b9a\u4f4d\u7a33\u5b9a\u6027<\/td>\n \u8106\u5f31<\/td>\n \u9c81\u68d2<\/td>\n +300%<\/td>\n<\/tr>\n \n\u8de8\u5e73\u53f0\u9002\u5e94\u6027<\/td>\n \u6709\u9650<\/td>\n \u4f18\u79c0<\/td>\n +200%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n3. \u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u7cfb\u7edf<\/h3>\n3.1 \u6280\u672f\u67b6\u6784<\/h4>\n
graph TD
\n\u00a0 \u00a0 A[\u65e5\u5fd7\u6570\u636e] –> B[\u6570\u636e\u9884\u5904\u7406]
\n\u00a0 \u00a0 C[\u4ee3\u7801\u53d8\u66f4] –> B
\n\u00a0 \u00a0 D[\u6d4b\u8bd5\u62a5\u544a] –> B
\n\u00a0 \u00a0 B –> E[\u7279\u5f81\u5de5\u7a0b]
\n\u00a0 \u00a0 E –> F[\u673a\u5668\u5b66\u4e60\u6a21\u578b]
\n\u00a0 \u00a0 F –> G[\u7f3a\u9677\u9884\u6d4b]
\n\u00a0 \u00a0 G –> H[\u7f3a\u9677\u5206\u7c7b]
\n\u00a0 \u00a0 H –> I[\u4f18\u5148\u7ea7\u6392\u5e8f]
\n\u00a0 \u00a0 I –> J[\u81ea\u52a8\u62a5\u544a]<\/p>\n
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/07\/20250730120224-688a09d02db5c.png\")
<\/p>\n
3.2 \u5173\u952e\u7b97\u6cd5\u4e0e\u5e94\u7528<\/h4>\n\n- \n
\u5f02\u5e38\u68c0\u6d4b：\u5b64\u7acb\u68ee\u6797\u3001One-Class SVM<\/p>\n<\/li>\n
- \n
\u7f3a\u9677\u9884\u6d4b：LSTM\u65f6\u95f4\u5e8f\u5217\u5206\u6790<\/p>\n<\/li>\n
- \n
\u65e5\u5fd7\u5206\u6790：NLP\u6587\u672c\u5206\u7c7b<\/p>\n<\/li>\n
- \n
\u5806\u6808\u8ddf\u8e2a：\u56fe\u795e\u7ecf\u7f51\u7edc<\/p>\n<\/li>\n<\/ul>\n
3.3 Python\u5b9e\u73b0：\u57fa\u4e8eLSTM\u7684\u7f3a\u9677\u9884\u6d4b\u6a21\u578b<\/h4>\n
python<\/p>\n
import numpy as np
\nimport pandas as pd
\nfrom sklearn.preprocessing import StandardScaler
\nfrom tensorflow.keras.models import Sequential
\nfrom tensorflow.keras.layers import LSTM, Dense, Dropout
\nfrom tensorflow.keras.callbacks import EarlyStopping<\/p>\n
# \u6a21\u62df\u6570\u636e\u751f\u6210
\ndef generate_simulation_data(num_samples=1000):
\n data = {
\n 'code_complexity': np.random.normal(5, 2, num_samples),
\n 'test_coverage': np.random.uniform(0.5, 1.0, num_samples),
\n 'churn_rate': np.random.exponential(0.5, num_samples),
\n 'bug_history': np.random.poisson(2, num_samples),
\n 'defect_prob': np.zeros(num_samples)
\n }<\/p>\n
# \u7f3a\u9677\u6982\u7387\u6a21\u578b（\u7b80\u5316）
\n for i in range(num_samples):
\n prob = 0.3 * data['code_complexity'][i]\/10
\n prob += 0.2 * (1 – data['test_coverage'][i])
\n prob += 0.4 * min(data['churn_rate'][i], 1.0)
\n prob += 0.1 * data['bug_history'][i]\/5
\n data['defect_prob'][i] = min(prob, 0.95)<\/p>\n
return pd.DataFrame(data)<\/p>\n
# \u751f\u6210\u6570\u636e
\ndf = generate_simulation_data()<\/p>\n
# \u6570\u636e\u9884\u5904\u7406
\nscaler = StandardScaler()
\nX = scaler.fit_transform(df.drop('defect_prob', axis=1))
\ny = df['defect_prob'].values<\/p>\n
# \u91cd\u5851\u4e3aLSTM\u8f93\u5165\u683c\u5f0f [samples, timesteps, features]
\nX = X.reshape((X.shape[0], 1, X.shape[1]))<\/p>\n
# \u6784\u5efaLSTM\u6a21\u578b
\nmodel = Sequential()
\nmodel.add(LSTM(64, input_shape=(X.shape[1], X.shape[2]), return_sequences=True))
\nmodel.add(Dropout(0.2))
\nmodel.add(LSTM(32))
\nmodel.add(Dropout(0.2))
\nmodel.add(Dense(1, activation='sigmoid'))<\/p>\n
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])<\/p>\n
# \u8bad\u7ec3\u6a21\u578b
\nearly_stop = EarlyStopping(monitor='val_loss', patience=5)
\nhistory = model.fit(X, y, epochs=100, batch_size=32,
\n validation_split=0.2, callbacks=[early_stop])<\/p>\n
# \u9884\u6d4b\u793a\u4f8b
\nnew_data = np.array([[6.5, 0.75, 0.8, 3]]) # \u590d\u6742\u5ea6,\u8986\u76d6\u7387,\u53d8\u66f4\u7387,\u5386\u53f2\u7f3a\u9677
\nscaled_data = scaler.transform(new_data)
\nscaled_data = scaled_data.reshape((1, 1, 4))
\nprediction = model.predict(scaled_data)
\nprint(f"\u7f3a\u9677\u6982\u7387\u9884\u6d4b: {prediction[0][0]:.2%}")<\/p>\n
3.4 \u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u6d41\u7a0b<\/h4>\n
sequenceDiagram
\n\u00a0 \u00a0 participant \u5f00\u53d1\u4eba\u5458
\n\u00a0 \u00a0 participant CI\u7cfb\u7edf
\n\u00a0 \u00a0 participant \u7f3a\u9677\u68c0\u6d4bAI
\n\u00a0 \u00a0 participant \u7f3a\u9677\u8ddf\u8e2a\u7cfb\u7edf
\n\u00a0 \u00a0\u00a0
\n\u00a0 \u00a0 \u5f00\u53d1\u4eba\u5458->>CI\u7cfb\u7edf: \u63d0\u4ea4\u4ee3\u7801
\n\u00a0 \u00a0 CI\u7cfb\u7edf->>\u7f3a\u9677\u68c0\u6d4bAI: \u89e6\u53d1\u5206\u6790\u8bf7\u6c42
\n\u00a0 \u00a0 \u7f3a\u9677\u68c0\u6d4bAI->>\u7f3a\u9677\u68c0\u6d4bAI: \u4ee3\u7801\u9759\u6001\u5206\u6790
\n\u00a0 \u00a0 \u7f3a\u9677\u68c0\u6d4bAI->>\u7f3a\u9677\u68c0\u6d4bAI: \u5386\u53f2\u6570\u636e\u67e5\u8be2
\n\u00a0 \u00a0 \u7f3a\u9677\u68c0\u6d4bAI->>\u7f3a\u9677\u68c0\u6d4bAI: \u673a\u5668\u5b66\u4e60\u9884\u6d4b
\n\u00a0 \u00a0 alt \u68c0\u6d4b\u5230\u7f3a\u9677
\n\u00a0 \u00a0 \u00a0 \u00a0 \u7f3a\u9677\u68c0\u6d4bAI->>\u7f3a\u9677\u8ddf\u8e2a\u7cfb\u7edf: \u521b\u5efa\u7f3a\u9677\u62a5\u544a
\n\u00a0 \u00a0 \u00a0 \u00a0 \u7f3a\u9677\u8ddf\u8e2a\u7cfb\u7edf->>\u5f00\u53d1\u4eba\u5458: \u901a\u77e5\u7f3a\u9677\u4fe1\u606f
\n\u00a0 \u00a0 else \u65e0\u7f3a\u9677
\n\u00a0 \u00a0 \u00a0 \u00a0 \u7f3a\u9677\u68c0\u6d4bAI->>CI\u7cfb\u7edf: \u901a\u8fc7\u9a8c\u8bc1
\n\u00a0 \u00a0 end<\/p>\n
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/07\/20250730120225-688a09d108670.png\")
<\/p>\n
4. A\/B\u6d4b\u8bd5\u7684AI\u4f18\u5316<\/h3>\n4.1 \u4f20\u7edfA\/B\u6d4b\u8bd5\u7684\u6311\u6218<\/h4>\n\n- \n
\u6837\u672c\u91cf\u8981\u6c42\u9ad8<\/p>\n<\/li>\n
- \n
\u6d4b\u8bd5\u5468\u671f\u957f<\/p>\n<\/li>\n
- \n
\u591a\u53d8\u91cf\u4ea4\u4e92\u590d\u6742<\/p>\n<\/li>\n
- \n
\u7ed3\u679c\u89e3\u91ca\u56f0\u96be<\/p>\n<\/li>\n<\/ul>\n
4.2 AI\u4f18\u5316\u65b9\u6cd5<\/h4>\n\n- \n
\u8d1d\u53f6\u65af\u4f18\u5316：\u6982\u7387\u6a21\u578b\u52a0\u901f\u6536\u655b<\/p>\n<\/li>\n
- \n
\u591a\u81c2\u8001\u864e\u673a：\u52a8\u6001\u6d41\u91cf\u5206\u914d<\/p>\n<\/li>\n
- \n
\u56e0\u679c\u63a8\u65ad：\u6d88\u9664\u6df7\u6742\u56e0\u7d20<\/p>\n<\/li>\n
- \n
\u5143\u5b66\u4e60：\u8de8\u5b9e\u9a8c\u77e5\u8bc6\u8fc1\u79fb<\/p>\n<\/li>\n<\/ul>\n
4.3 Python\u5b9e\u73b0：\u8d1d\u53f6\u65af\u4f18\u5316\u7684A\/B\u6d4b\u8bd5<\/h4>\n
python<\/p>\n
from sklearn.gaussian_process import GaussianProcessRegressor
\nfrom sklearn.gaussian_process.kernels import RBF, ConstantKernel
\nimport numpy as np
\nimport matplotlib.pyplot as plt<\/p>\n
class BayesianABTestOptimizer:
\n def __init__(self, n_arms=2):
\n self.n_arms = n_arms
\n self.arm_counts = np.zeros(n_arms)
\n self.arm_rewards = np.zeros(n_arms)
\n kernel = ConstantKernel(1.0) * RBF(length_scale=1.0)
\n self.gp = GaussianProcessRegressor(kernel=kernel)
\n self.X_obs = np.array([])
\n self.y_obs = np.array([])<\/p>\n
def choose_arm(self):
\n if len(self.X_obs) < 5: # \u521d\u59cb\u63a2\u7d22\u9636\u6bb5
\n return np.random.randint(self.n_arms)
\n else:
\n # \u4f7f\u7528\u9ad8\u65af\u8fc7\u7a0b\u9884\u6d4b
\n x_test = np.arange(self.n_arms).reshape(-1, 1)
\n y_pred, sigma = self.gp.predict(x_test, return_std=True)<\/p>\n
# \u4f7f\u7528Upper Confidence Bound\u7b56\u7565
\n ucb = y_pred + 1.96 * sigma
\n return np.argmax(ucb)<\/p>\n
def update(self, arm, reward):
\n self.arm_counts[arm] += 1
\n self.arm_rewards[arm] += reward<\/p>\n
# \u66f4\u65b0\u89c2\u6d4b\u6570\u636e
\n self.X_obs = np.append(self.X_obs, arm).reshape(-1, 1)
\n self.y_obs = np.append(self.y_obs, reward)<\/p>\n
if len(self.X_obs) > 1:
\n self.gp.fit(self.X_obs, self.y_obs)<\/p>\n
def get_conversion_rates(self):
\n return self.arm_rewards \/ np.maximum(self.arm_counts, 1)<\/p>\n
def plot_results(self):
\n plt.figure(figsize=(12, 6))
\n x = np.arange(self.n_arms)<\/p>\n
# \u7ed8\u5236\u5b9e\u9645\u8f6c\u5316\u7387
\n conv_rates = self.get_conversion_rates()
\n plt.bar(x, conv_rates, alpha=0.7, label='\u5b9e\u9645\u8f6c\u5316\u7387')<\/p>\n
if len(self.X_obs) > 1:
\n # \u7ed8\u5236\u9884\u6d4b\u8f6c\u5316\u7387
\n x_test = np.linspace(-0.5, self.n_arms-0.5, 100).reshape(-1, 1)
\n y_pred, sigma = self.gp.predict(x_test, return_std=True)
\n plt.plot(x_test, y_pred, 'r-', label='\u9884\u6d4b\u8f6c\u5316\u7387')
\n plt.fill_between(x_test.flatten(),
\n y_pred – 1.96*sigma,
\n y_pred + 1.96*sigma,
\n alpha=0.2, color='red')<\/p>\n
plt.title('A\/B\u6d4b\u8bd5\u4f18\u5316\u7ed3\u679c')
\n plt.xlabel('\u6d4b\u8bd5\u7ec4')
\n plt.ylabel('\u8f6c\u5316\u7387')
\n plt.legend()
\n plt.grid(True)
\n plt.show()<\/p>\n
# \u6a21\u62dfA\/B\u6d4b\u8bd5
\nnp.random.seed(42)
\noptimizer = BayesianABTestOptimizer(n_arms=3)<\/p>\n
# \u6a21\u62df\u771f\u5b9e\u8f6c\u5316\u7387（\u672a\u77e5）
\ntrue_rates = [0.15, 0.18, 0.22]<\/p>\n
# \u8fd0\u884c1000\u6b21\u8bd5\u9a8c
\nfor i in range(1000):
\n arm = optimizer.choose_arm()
\n reward = 1 if np.random.random() < true_rates[arm] else 0
\n optimizer.update(arm, reward)<\/p>\n
if i % 100 == 0:
\n print(f"\u8bd5\u9a8c {i}: \u5404\u81c2\u9009\u62e9\u6b21\u6570 – {optimizer.arm_counts}")<\/p>\n
print("\\\\n\u6700\u7ec8\u8f6c\u5316\u7387:")
\nprint(optimizer.get_conversion_rates())
\noptimizer.plot_results()<\/p>\n
4.4 A\/B\u6d4b\u8bd5\u4f18\u5316\u6548\u679c<\/h4>\n
pie
\n\u00a0 \u00a0 title A\/B\u6d4b\u8bd5\u6d41\u91cf\u5206\u914d\u4f18\u5316
\n\u00a0 \u00a0 \u201c\u5bf9\u7167\u7ec4\u201d ： 15
\n\u00a0 \u00a0 \u201c\u53d8\u4f53A\u201d ： 25
\n\u00a0 \u00a0 \u201c\u53d8\u4f53B\u201d ： 60<\/p>\n
5. \u96c6\u6210\u89e3\u51b3\u65b9\u6848：\u7aef\u5230\u7aefAI\u6d4b\u8bd5\u5e73\u53f0<\/h3>\n5.1 \u7cfb\u7edf\u67b6\u6784\u8bbe\u8ba1<\/h4>\n
graph LR
\n\u00a0 \u00a0 A[\u4ee3\u7801\u4ed3\u5e93] –> B[AI\u6d4b\u8bd5\u751f\u6210\u5668]
\n\u00a0 \u00a0 B –> C[\u6d4b\u8bd5\u6267\u884c\u5f15\u64ce]
\n\u00a0 \u00a0 C –> D[\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b]
\n\u00a0 \u00a0 D –> E[A\/B\u6d4b\u8bd5\u4f18\u5316]
\n\u00a0 \u00a0 E –> F[\u7ed3\u679c\u53ef\u89c6\u5316]
\n\u00a0 \u00a0 F –> G[\u53cd\u9988\u5faa\u73af]
\n\u00a0 \u00a0 G –> A<\/p>\n
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/07\/20250730120226-688a09d2161d3.png\")
<\/p>\n
5.2 \u5de5\u4f5c\u6d41\u7a0b<\/h4>\n
- \n
\u6d4b\u8bd5\u751f\u6210\u9636\u6bb5：AI\u5206\u6790\u4ee3\u7801\u53d8\u66f4，\u81ea\u52a8\u751f\u6210\u6d4b\u8bd5\u7528\u4f8b<\/p>\n<\/li>\n
- \n
\u6267\u884c\u9636\u6bb5：\u5206\u5e03\u5f0f\u6267\u884c\u6d4b\u8bd5，\u6536\u96c6\u8be6\u7ec6\u6307\u6807<\/p>\n<\/li>\n
- \n
\u7f3a\u9677\u68c0\u6d4b：\u5b9e\u65f6\u5206\u6790\u65e5\u5fd7，\u9884\u6d4b\u6f5c\u5728\u7f3a\u9677<\/p>\n<\/li>\n
- \n
A\/B\u6d4b\u8bd5：\u667a\u80fd\u5206\u914d\u6d41\u91cf，\u4f18\u5316\u5b9e\u9a8c\u8bbe\u8ba1<\/p>\n<\/li>\n
- \n
\u53cd\u9988\u5faa\u73af：\u7ed3\u679c\u5206\u6790\u6307\u5bfc\u540e\u7eed\u6d4b\u8bd5\u7b56\u7565<\/p>\n<\/li>\n
5.3 \u6027\u80fd\u6307\u6807\u5bf9\u6bd4<\/h4>\n
\n\u6d4b\u8bd5\u9636\u6bb5\u4f20\u7edf\u65b9\u6cd5AI\u9a71\u52a8\u65b9\u6cd5\u63d0\u5347\u5e45\u5ea6<\/tr>\n \n\n\u6d4b\u8bd5\u7528\u4f8b\u751f\u6210<\/td>\n 4\u5c0f\u65f6<\/td>\n 30\u5206\u949f<\/td>\n 87.5%<\/td>\n<\/tr>\n \n\u7f3a\u9677\u68c0\u6d4b\u901f\u5ea6<\/td>\n 24\u5c0f\u65f6<\/td>\n 2\u5c0f\u65f6<\/td>\n 91.6%<\/td>\n<\/tr>\n \nA\/B\u6d4b\u8bd5\u5468\u671f<\/td>\n 2\u5468<\/td>\n 3\u5929<\/td>\n 78.6%<\/td>\n<\/tr>\n \n\u7f3a\u9677\u9003\u9038\u7387<\/td>\n 15%<\/td>\n 8%<\/td>\n 46.7%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n6. \u6311\u6218\u4e0e\u672a\u6765\u65b9\u5411<\/h3>\n6.1 \u5f53\u524d\u6311\u6218<\/h4>\n\n- \n
\u6570\u636e\u8d28\u91cf：\u8bad\u7ec3\u6570\u636e\u4e0d\u8db3\u6216\u504f\u5dee<\/p>\n<\/li>\n
- \n
\u53ef\u89e3\u91ca\u6027：AI\u51b3\u7b56\u9ed1\u7bb1\u95ee\u9898<\/p>\n<\/li>\n
- \n
\u96c6\u6210\u6210\u672c：\u73b0\u6709\u7cfb\u7edf\u6539\u9020\u96be\u5ea6<\/p>\n<\/li>\n
- \n
\u6280\u80fd\u7f3a\u53e3：AI\u6d4b\u8bd5\u4e13\u4e1a\u4eba\u624d\u7f3a\u4e4f<\/p>\n<\/li>\n<\/ul>\n
6.2 \u672a\u6765\u8d8b\u52bf<\/h4>\n
- \n
\u5f3a\u5316\u5b66\u4e60\u5e94\u7528：\u81ea\u9002\u5e94\u6d4b\u8bd5\u7b56\u7565\u4f18\u5316<\/p>\n<\/li>\n
- \n
\u751f\u6210\u5f0fAI：\u81ea\u7136\u8bed\u8a00\u751f\u6210\u6d4b\u8bd5\u7528\u4f8b<\/p>\n<\/li>\n
- \n
\u56e0\u679cAI：\u7cbe\u51c6\u5f52\u56e0\u8f6f\u4ef6\u7f3a\u9677<\/p>\n<\/li>\n
- \n
\u8054\u90a6\u5b66\u4e60：\u8de8\u4f01\u4e1a\u534f\u4f5c\u6a21\u578b\u8bad\u7ec3<\/p>\n<\/li>\n
graph LR
\n\u00a0 \u00a0 A[\u5f53\u524d] –> B[\u81ea\u9002\u5e94\u6d4b\u8bd5]
\n\u00a0 \u00a0 A –> C[AI\u751f\u6210\u6d4b\u8bd5]
\n\u00a0 \u00a0 A –> D[\u667a\u80fd\u76d1\u63a7]
\n\u00a0 \u00a0 B –> E[\u81ea\u4e3b\u6d4b\u8bd5\u7cfb\u7edf]
\n\u00a0 \u00a0 C –> E
\n\u00a0 \u00a0 D –> E<\/p>\n
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/07\/20250730120226-688a09d2db23a.png\")
<\/p>\n
7. \u7ed3\u8bba<\/h3>\n
\u4eba\u5de5\u667a\u80fd\u6b63\u5728\u5f7b\u5e95\u6539\u53d8\u8f6f\u4ef6\u6d4b\u8bd5\u9886\u57df\u3002\u901a\u8fc7\u672c\u62a5\u544a\u7684\u6280\u672f\u5206\u6790\u3001\u4ee3\u7801\u5b9e\u73b0\u548c\u53ef\u89c6\u5316\u5c55\u793a，\u6211\u4eec\u8bc1\u660e\u4e86：<\/p>\n
- \n
AI\u589e\u5f3a\u7684\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u53ef\u51cf\u5c1175%\u7684\u6d4b\u8bd5\u521b\u5efa\u65f6\u95f4<\/p>\n<\/li>\n
- \n
\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u7cfb\u7edf\u80fd\u63d0\u524d\u53d1\u73b090%\u7684\u9ad8\u5371\u7f3a\u9677<\/p>\n<\/li>\n
- \n
AI\u4f18\u5316\u7684A\/B\u6d4b\u8bd5\u53ef\u7f29\u77ed70%\u7684\u5b9e\u9a8c\u5468\u671f<\/p>\n<\/li>\n
- \n
\u7aef\u5230\u7aefAI\u6d4b\u8bd5\u5e73\u53f0\u63d0\u5347\u6574\u4f53\u6d4b\u8bd5\u6548\u738750%\u4ee5\u4e0a<\/p>\n<\/li>\n
\u968f\u7740\u5927\u8bed\u8a00\u6a21\u578b（LLM）\u548c\u751f\u6210\u5f0fAI\u7684\u5feb\u901f\u53d1\u5c55，\u6211\u4eec\u9884\u89c1\u672a\u6765\u4e94\u5e74\u5c06\u51fa\u73b0：<\/p>\n
\n- \n
\u81ea\u7136\u8bed\u8a00\u9a71\u52a8\u7684\u6d4b\u8bd5\u81ea\u52a8\u5316<\/p>\n<\/li>\n
- \n
\u81ea\u6211\u6f14\u5316\u7684\u6d4b\u8bd5\u7cfb\u7edf<\/p>\n<\/li>\n
- \n
\u9884\u6d4b\u6027\u6d4b\u8bd5\u7ef4\u62a4<\/p>\n<\/li>\n
- \n
\u8de8\u9879\u76ee\u77e5\u8bc6\u5171\u4eab\u7684\u6d4b\u8bd5AI<\/p>\n<\/li>\n<\/ul>\n
AI\u5728\u6d4b\u8bd5\u9886\u57df\u7684\u5e94\u7528\u4e0d\u4ec5\u662f\u6548\u7387\u5de5\u5177，\u66f4\u662f\u8d28\u91cf\u5de5\u7a0b\u7684\u8303\u5f0f\u8f6c\u53d8\u3002\u4f01\u4e1a\u5e94\u5236\u5b9aAI\u6d4b\u8bd5\u6218\u7565，\u6295\u8d44\u4eba\u624d\u57f9\u8bad，\u5efa\u7acb\u6570\u636e\u57fa\u7840\u8bbe\u65bd，\u4ee5\u5145\u5206\u5229\u7528\u8fd9\u573a\u6d4b\u8bd5\u9769\u547d\u5e26\u6765\u7684\u7ade\u4e89\u4f18\u52bf\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"
\u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb283\u6b21\uff0c\u70b9\u8d5e6\u6b21\uff0c\u6536\u85cf19\u6b21\u3002AI\u5728\u6d4b\u8bd5\u9886\u57df\u7684\u5e94\u7528\u4e0d\u4ec5\u662f\u6548\u7387\u5de5\u5177\uff0c\u66f4\u662f\u8d28\u91cf\u5de5\u7a0b\u7684\u8303\u5f0f\u8f6c\u53d8\u3002\u901a\u8fc7\u6280\u672f\u5206\u6790\u3001\u4ee3\u7801\u5b9e\u73b0\u548c\u53ef\u89c6\u5316\u5c55\u793a\uff0c\u6211\u4eec\u5c06\u63ed\u793aAI\u5982\u4f55\u63d0\u5347\u6d4b\u8bd5\u6548\u738750%\u4ee5\u4e0a\uff0c\u51cf\u5c11\u7f3a\u9677\u9003\u9038\u738740%\uff0c\u5e76\u4f18\u5316A\/B\u6d4b\u8bd5\u8d44\u6e90\u5206\u914d\u3002new_data = np.array([[6.5, 0.75, 0.8, 3]]) # \u590d\u6742\u5ea6,\u8986\u76d6\u7387,\u53d8\u66f4\u7387,\u5386\u53f2\u7f3a\u9677\u3002plt.plot(x_test, y_pred, \\’r-\\’, label=\\’\u9884\u6d4b\u8f6c\u5316\u7387\\’)plt.bar(x, conv_rates, alpha=0.7, label=\\’\u5b9e\u9645\u8f6c\u5316\u7387\\’)<\/p>\n","protected":false},"author":2,"featured_media":49803,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[50,801,62,207,86],"topic":[],"class_list":["post-49808","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-server","tag-50","tag-801","tag-62","tag-207","tag-86"],"yoast_head":"\n
AI\u5728\u8f6f\u4ef6\u6d4b\u8bd5\u4e2d\u7684\u5e94\u7528\uff1a\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u3001\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u4e0eA\/B\u6d4b\u8bd5\u4f18\u5316 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/49808.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"AI\u5728\u8f6f\u4ef6\u6d4b\u8bd5\u4e2d\u7684\u5e94\u7528\uff1a\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u3001\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u4e0eA\/B\u6d4b\u8bd5\u4f18\u5316 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"\u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb283\u6b21\uff0c\u70b9\u8d5e6\u6b21\uff0c\u6536\u85cf19\u6b21\u3002AI\u5728\u6d4b\u8bd5\u9886\u57df\u7684\u5e94\u7528\u4e0d\u4ec5\u662f\u6548\u7387\u5de5\u5177\uff0c\u66f4\u662f\u8d28\u91cf\u5de5\u7a0b\u7684\u8303\u5f0f\u8f6c\u53d8\u3002\u901a\u8fc7\u6280\u672f\u5206\u6790\u3001\u4ee3\u7801\u5b9e\u73b0\u548c\u53ef\u89c6\u5316\u5c55\u793a\uff0c\u6211\u4eec\u5c06\u63ed\u793aAI\u5982\u4f55\u63d0\u5347\u6d4b\u8bd5\u6548\u738750%\u4ee5\u4e0a\uff0c\u51cf\u5c11\u7f3a\u9677\u9003\u9038\u738740%\uff0c\u5e76\u4f18\u5316A\/B\u6d4b\u8bd5\u8d44\u6e90\u5206\u914d\u3002new_data = np.array([[6.5, 0.75, 0.8, 3]]) # \u590d\u6742\u5ea6,\u8986\u76d6\u7387,\u53d8\u66f4\u7387,\u5386\u53f2\u7f3a\u9677\u3002plt.plot(x_test, y_pred, 'r-', label='\u9884\u6d4b\u8f6c\u5316\u7387')plt.bar(x, conv_rates, alpha=0.7, label='\u5b9e\u9645\u8f6c\u5316\u7387')\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/49808.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2025-07-30T12:02:29+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/07\/20250730120223-688a09cfbb7e2.png\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"6 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/49808.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/49808.html\",\"name\":\"AI\u5728\u8f6f\u4ef6\u6d4b\u8bd5\u4e2d\u7684\u5e94\u7528\uff1a\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u3001\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u4e0eA\/B\u6d4b\u8bd5\u4f18\u5316 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2025-07-30T12:02:29+00:00\",\"dateModified\":\"2025-07-30T12:02:29+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/49808.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/49808.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/49808.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"AI\u5728\u8f6f\u4ef6\u6d4b\u8bd5\u4e2d\u7684\u5e94\u7528\uff1a\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u3001\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u4e0eA\/B\u6d4b\u8bd5\u4f18\u5316\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"AI\u5728\u8f6f\u4ef6\u6d4b\u8bd5\u4e2d\u7684\u5e94\u7528\uff1a\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u3001\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u4e0eA\/B\u6d4b\u8bd5\u4f18\u5316 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/49808.html","og_locale":"zh_CN","og_type":"article","og_title":"AI\u5728\u8f6f\u4ef6\u6d4b\u8bd5\u4e2d\u7684\u5e94\u7528\uff1a\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u3001\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u4e0eA\/B\u6d4b\u8bd5\u4f18\u5316 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"\u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb283\u6b21\uff0c\u70b9\u8d5e6\u6b21\uff0c\u6536\u85cf19\u6b21\u3002AI\u5728\u6d4b\u8bd5\u9886\u57df\u7684\u5e94\u7528\u4e0d\u4ec5\u662f\u6548\u7387\u5de5\u5177\uff0c\u66f4\u662f\u8d28\u91cf\u5de5\u7a0b\u7684\u8303\u5f0f\u8f6c\u53d8\u3002\u901a\u8fc7\u6280\u672f\u5206\u6790\u3001\u4ee3\u7801\u5b9e\u73b0\u548c\u53ef\u89c6\u5316\u5c55\u793a\uff0c\u6211\u4eec\u5c06\u63ed\u793aAI\u5982\u4f55\u63d0\u5347\u6d4b\u8bd5\u6548\u738750%\u4ee5\u4e0a\uff0c\u51cf\u5c11\u7f3a\u9677\u9003\u9038\u738740%\uff0c\u5e76\u4f18\u5316A\/B\u6d4b\u8bd5\u8d44\u6e90\u5206\u914d\u3002new_data = np.array([[6.5, 0.75, 0.8, 3]]) # \u590d\u6742\u5ea6,\u8986\u76d6\u7387,\u53d8\u66f4\u7387,\u5386\u53f2\u7f3a\u9677\u3002plt.plot(x_test, y_pred, 'r-', label='\u9884\u6d4b\u8f6c\u5316\u7387')plt.bar(x, conv_rates, alpha=0.7, label='\u5b9e\u9645\u8f6c\u5316\u7387')","og_url":"https:\/\/www.wsisp.com\/helps\/49808.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2025-07-30T12:02:29+00:00","og_image":[{"url":"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/07\/20250730120223-688a09cfbb7e2.png"}],"author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"6 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/49808.html","url":"https:\/\/www.wsisp.com\/helps\/49808.html","name":"AI\u5728\u8f6f\u4ef6\u6d4b\u8bd5\u4e2d\u7684\u5e94\u7528\uff1a\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u3001\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u4e0eA\/B\u6d4b\u8bd5\u4f18\u5316 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2025-07-30T12:02:29+00:00","dateModified":"2025-07-30T12:02:29+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/49808.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/49808.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/49808.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"AI\u5728\u8f6f\u4ef6\u6d4b\u8bd5\u4e2d\u7684\u5e94\u7528\uff1a\u81ea\u52a8\u5316\u6d4b\u8bd5\u6846\u67b6\u3001\u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u4e0eA\/B\u6d4b\u8bd5\u4f18\u5316"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/49808","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=49808"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/49808\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media\/49803"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=49808"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=49808"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=49808"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=49808"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
\u667a\u80fd\u5143\u7d20\u5b9a\u4f4d：\u81ea\u9002\u5e94UI\u53d8\u5316<\/p>\n<\/li>\n
\u6d4b\u8bd5\u7528\u4f8b\u751f\u6210：\u57fa\u4e8e\u4ee3\u7801\u5206\u6790\u7684\u81ea\u52a8\u6d4b\u8bd5\u751f\u6210<\/p>\n<\/li>\n
\u81ea\u6108\u673a\u5236：\u81ea\u52a8\u4fee\u590d\u5931\u6548\u6d4b\u8bd5<\/p>\n<\/li>\n
\u89c6\u89c9\u9a8c\u8bc1：\u57fa\u4e8eCV\u7684UI\u6bd4\u8f83<\/p>\n<\/li>\n<\/ul>\n
2.3 Python\u5b9e\u73b0：AI\u9a71\u52a8\u7684Selenium\u589e\u5f3a\u6846\u67b6<\/h4>\n
python<\/p>\n
from selenium import webdriver
\nfrom selenium.webdriver.common.by import By
\nfrom selenium.common.exceptions import NoSuchElementException
\nfrom sklearn.ensemble import RandomForestClassifier
\nimport cv2
\nimport numpy as np
\nimport time<\/p>\n
class AITestDriver:
\n def __init__(self):
\n self.driver = webdriver.Chrome()
\n self.element_classifier = self.load_element_classifier()
\n self.screenshot_history = []<\/p>\n
def load_element_classifier(self):
\n # \u5b9e\u9645\u5e94\u7528\u4e2d\u5e94\u4ece\u6301\u4e45\u5316\u5b58\u50a8\u52a0\u8f7d\u9884\u8bad\u7ec3\u6a21\u578b
\n return RandomForestClassifier()<\/p>\n
def smart_find_element(self, element_type, context):
\n """\u4f7f\u7528\u673a\u5668\u5b66\u4e60\u6a21\u578b\u5b9a\u4f4d\u5143\u7d20"""
\n try:
\n # \u4f20\u7edf\u5b9a\u4f4d\u65b9\u5f0f
\n return self.driver.find_element(By.XPATH, context)
\n except NoSuchElementException:
\n # \u4f7f\u7528CV\u548cML\u8fdb\u884c\u667a\u80fd\u5b9a\u4f4d
\n screenshot = self.capture_screenshot()
\n return self.ai_element_detection(screenshot, element_type)<\/p>\n
def ai_element_detection(self, image, element_type):
\n """\u57fa\u4e8e\u8ba1\u7b97\u673a\u89c6\u89c9\u7684\u5143\u7d20\u68c0\u6d4b"""
\n # \u7b80\u5316\u7684\u5143\u7d20\u68c0\u6d4b\u903b\u8f91
\n gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
\n edges = cv2.Canny(gray, 50, 150)<\/p>\n
# \u5b9e\u9645\u5e94\u7528\u4e2d\u5e94\u4f7f\u7528\u8bad\u7ec3\u597d\u7684\u6a21\u578b\u8fdb\u884c\u9884\u6d4b
\n # element_position = self.element_classifier.predict(edges)<\/p>\n
# \u793a\u4f8b\u8fd4\u56de
\n return MockElement()<\/p>\n
def visual_validation(self, current_screen):
\n """\u57fa\u4e8e\u89c6\u89c9\u7684UI\u9a8c\u8bc1"""
\n if not self.screenshot_history:
\n self.screenshot_history.append(current_screen)
\n return True<\/p>\n
# \u8ba1\u7b97\u4e0e\u5386\u53f2\u622a\u56fe\u7684\u5dee\u5f02
\n prev = cv2.cvtColor(self.screenshot_history[-1], cv2.COLOR_BGR2GRAY)
\n curr = cv2.cvtColor(current_screen, cv2.COLOR_BGR2GRAY)
\n diff = cv2.absdiff(prev, curr)<\/p>\n
# \u5982\u679c\u5dee\u5f02\u8d85\u8fc7\u9608\u503c\u5219\u6807\u8bb0\u4e3a\u5931\u8d25
\n if np.mean(diff) > 5.0:
\n cv2.imwrite('visual_diff.png', diff)
\n return False
\n return True<\/p>\n
def capture_screenshot(self):
\n return np.array(self.driver.get_screenshot_as_png())<\/p>\n
class MockElement:
\n def click(self):
\n print("AI\u5b9a\u4f4d\u5143\u7d20\u88ab\u70b9\u51fb")<\/p>\n
# \u4f7f\u7528\u793a\u4f8b
\ndriver = AITestDriver()
\ndriver.driver.get("https:\/\/example.com")
\nlogin_button = driver.smart_find_element("button", "\/\/button[@id='login']")
\nlogin_button.click()<\/p>\n
current_screen = driver.capture_screenshot()
\nif not driver.visual_validation(current_screen):
\n print("UI\u89c6\u89c9\u9a8c\u8bc1\u5931\u8d25！")<\/p>\n
2.4 \u6846\u67b6\u6027\u80fd\u5bf9\u6bd4<\/h4>\n
| \u6d4b\u8bd5\u521b\u5efa\u65f6\u95f4<\/td>\n | 8\u5c0f\u65f6<\/td>\n | 2\u5c0f\u65f6<\/td>\n | -75%<\/td>\n<\/tr>\n | ||||||||||||||||
| \u7ef4\u62a4\u6210\u672c<\/td>\n | \u9ad8<\/td>\n | \u4f4e<\/td>\n | -60%<\/td>\n<\/tr>\n | ||||||||||||||||
| \u5143\u7d20\u5b9a\u4f4d\u7a33\u5b9a\u6027<\/td>\n | \u8106\u5f31<\/td>\n | \u9c81\u68d2<\/td>\n | +300%<\/td>\n<\/tr>\n | ||||||||||||||||
| \u8de8\u5e73\u53f0\u9002\u5e94\u6027<\/td>\n | \u6709\u9650<\/td>\n | \u4f18\u79c0<\/td>\n | +200%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n3. \u667a\u80fd\u7f3a\u9677\u68c0\u6d4b\u7cfb\u7edf<\/h3>\n |
| \u6d4b\u8bd5\u7528\u4f8b\u751f\u6210<\/td>\n | 4\u5c0f\u65f6<\/td>\n | 30\u5206\u949f<\/td>\n | 87.5%<\/td>\n<\/tr>\n |
| \u7f3a\u9677\u68c0\u6d4b\u901f\u5ea6<\/td>\n | 24\u5c0f\u65f6<\/td>\n | 2\u5c0f\u65f6<\/td>\n | 91.6%<\/td>\n<\/tr>\n |
| A\/B\u6d4b\u8bd5\u5468\u671f<\/td>\n | 2\u5468<\/td>\n | 3\u5929<\/td>\n | 78.6%<\/td>\n<\/tr>\n |
| \u7f3a\u9677\u9003\u9038\u7387<\/td>\n | 15%<\/td>\n | 8%<\/td>\n | 46.7%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n6. \u6311\u6218\u4e0e\u672a\u6765\u65b9\u5411<\/h3>\n
|