import pandas as pd import numpy as np from sklearn.linear_model import LinearRegression

读取数据

data = pd.read_csv('test.csv')

数据预处理

data['销售日期'] = pd.to_datetime(data['销售日期']) data['扫码销售时间'] = pd.to_datetime(data['扫码销售时间']) data['销量(千克)'] = data['销量(千克)'].astype(float) data['销售单价(元/千克)'] = data['销售单价(元/千克)'].astype(float) data['是否打折销售'] = data['是否打折销售'].map({'是': 1, '否': 0}) data['批发价格'] = data['批发价格'].astype(float) data['单品损耗率'] = data['单品损耗率'].astype(float)

提取需要的特征

features = ['单品编码', '销售日期', '销售单价(元/千克)', '是否打折销售', '批发价格', '单品损耗率'] X = data[features]

提取目标变量

y = data['销量(千克)']

构建模型

model = LinearRegression() model.fit(X, y)

预测补货量和定价策略

n = 10 # 可售单品数量 x = np.zeros(n) # 补货量 p = np.zeros(n) # 定价策略

for i in range(n): x[i] = max(2.5, model.predict([[i, pd.to_datetime('2020-08-01'), 0, 0, 0, 0]])) p[i] = model.predict([[i, pd.to_datetime('2020-08-01'), 0, 0, 0, 0]])

计算商超收益

revenue = np.sum(x * p * (1 - data['单品损耗率']) - x * data['批发价格'])

print("补货量:", x) print("定价策略:", p) print("商超收益:", revenue)