以下是一个完整的Python代码，用于根据气象站历史降雨及其他天气特征数据集（`train.csv` 和 `test.csv`）进行明天降雨概率预测，并将结果保存到指定路径下的 `test_prediction.csv` 文件中，且不使用机器学习框架： ```python import csv import math import os # 加载数据 def load_data(file_path): data = [] with open(file_path, 'r', encoding='utf-8') as file: reader = csv.reader(file) headers = next(reader) for row in reader: # 处理缺失值，这里简单将非数字转为0 data.append([float(x) if x.replace('.', '', 1).isdigit() else 0 for x in row]) return headers, data # sigmoid函数 def sigmoid(z): return 1 / (1 + math.exp(-z)) # 训练逻辑回归模型 def train_logistic_regression(train_features, train_labels, learning_rate=0.01, iterations=1000): num_features = len(train_features[0]) weights = [0] * num_features bias = 0 for _ in range(iterations): for i in range(len(train_features)): z = sum([weights[j] * train_features[i][j] for j in range(num_features)]) + bias prediction = sigmoid(z) error = prediction - train_labels[i] for j in range(num_features): weights[j] -= learning_rate * error * train_features[i][j] bias -= learning_rate * error return weights, bias # 预测函数 def predict(test_features, weights, bias): predictions = [] for row in test_features: z = sum([weights[j] * row[j] for j in range(len(weights))]) + bias prediction = sigmoid(z) predictions.append(prediction) return predictions # 读取训练集和测试集 train_headers, train_data = load_data('input/train.csv') test_headers, test_data = load_data('input/test.csv') # 分离特征和标签 train_features = [row[2:-2] for row in train_data] # 去掉日期、地点、RainToday、RainTomorrow train_labels = [1 if row[-1] == 'Yes' else 0 for row in train_data] test_features = [row[2:-1] for row in test_data] # 去掉日期、地点、RainTomorrow # 训练模型 weights, bias = train_logistic_regression(train_features, train_labels) # 进行预测 predictions = predict(test_features, weights, bias) # 创建输出目录 output_dir = './output' if not os.path.exists(output_dir): os.makedirs(output_dir) # 保存预测结果 output_file = os.path.join(output_dir, 'test_prediction.csv') with open(output_file, 'w', encoding='utf-8', newline='') as file: writer = csv.writer(file) writer.writerow(['Date', 'Location', 'RainTomorrow']) for i in range(len(test_data)): writer.writerow([test_data[i][0], test_data[i][1], predictions[i]]) ```