当前位置：首页 > news >正文

Python实现机器学习驱动的智能医疗预测模型系统的示例代码框架

news 2025/7/8 10:35:12

以下是一个使用Python实现机器学习驱动的智能医疗预测模型系统的示例代码框架。这个框架涵盖了数据收集（爬虫）、数据清洗和预处理、模型构建（决策树和神经网络）以及模型评估的主要步骤。

1. 数据收集（爬虫）

首先，我们需要从网站获取X光影像数据。假设我们要爬取的网站允许爬取，并且遵循相关法律法规。这里我们使用requests和BeautifulSoup库来进行网页数据的抓取。

import requests
from bs4 import BeautifulSoup
import osdef download_xray_images(url, save_dir):if not os.path.exists(save_dir):os.makedirs(save_dir)response = requests.get(url)soup = BeautifulSoup(response.content, 'html.parser')image_tags = soup.find_all('img')for index, img in enumerate(image_tags):img_url = img.get('src')if img_url and (img_url.endswith('.jpg') or img_url.endswith('.jpeg') or img_url.endswith('.png')):img_response = requests.get(img_url)with open(os.path.join(save_dir, f'image_{index}.jpg'), 'wb') as f:f.write(img_response.content)

2. 数据清洗和预处理

接下来，我们需要对获取的X光影像数据进行清洗和预处理。这包括图像的读取、调整大小、归一化等操作。我们使用Pillow和numpy库来处理图像数据。

from PIL import Image
import numpy as npdef preprocess_images(image_dir, target_size=(224, 224)):images = []labels = []for root, dirs, files in os.walk(image_dir):for file in files:if file.endswith('.jpg') or file.endswith('.jpeg') or file.endswith('.png'):image_path = os.path.join(root, file)img = Image.open(image_path)img = img.resize(target_size)img = np.array(img)img = img / 255.0images.append(img)# 这里假设目录名就是标签label = os.path.basename(root)labels.append(label)return np.array(images), np.array(labels)

3. 特征提取

对于图像数据，我们可以使用预训练的卷积神经网络（如VGG16）来提取特征。

from keras.applications.vgg16 import VGG16, preprocess_input
from keras.models import Modeldef extract_features(images):base_model = VGG16(weights='imagenet', include_top=False)model = Model(inputs=base_model.input, outputs=base_model.output)images = preprocess_input(images)features = model.predict(images)features = features.flatten().reshape(features.shape[0], -1)return features

4. 模型构建

我们将使用决策树和神经网络模型进行疾病分类。

from sklearn.tree import DecisionTreeClassifier
from keras.models import Sequential
from keras.layers import Dense
from sklearn.model_selection import train_test_splitdef build_decision_tree_model(X, y):X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)model = DecisionTreeClassifier()model.fit(X_train, y_train)return model, X_test, y_testdef build_neural_network_model(X, y):X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)model = Sequential()model.add(Dense(128, activation='relu', input_shape=(X.shape[1],)))model.add(Dense(len(np.unique(y)), activation='softmax'))model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])model.fit(X_train, y_train, epochs=10, batch_size=32, verbose=1)return model, X_test, y_test

5. 模型评估

最后，我们需要评估模型的准确率和召回率。

from sklearn.metrics import accuracy_score, recall_scoredef evaluate_model(model, X_test, y_test):y_pred = model.predict(X_test)accuracy = accuracy_score(y_test, y_pred)recall = recall_score(y_test, y_pred, average='weighted')return accuracy, recall

主程序

if __name__ == "__main__":# 数据收集url = "your_target_url"save_dir = "xray_images"download_xray_images(url, save_dir)# 数据清洗和预处理images, labels = preprocess_images(save_dir)# 特征提取features = extract_features(images)# 决策树模型dt_model, dt_X_test, dt_y_test = build_decision_tree_model(features, labels)dt_accuracy, dt_recall = evaluate_model(dt_model, dt_X_test, dt_y_test)print(f"Decision Tree - Accuracy: {dt_accuracy}, Recall: {dt_recall}")# 神经网络模型nn_model, nn_X_test, nn_y_test = build_neural_network_model(features, labels)nn_accuracy, nn_recall = evaluate_model(nn_model, nn_X_test, nn_y_test)print(f"Neural Network - Accuracy: {nn_accuracy}, Recall: {nn_recall}")