【数据分析&数据挖掘】非数值型数据的哑变量转化、连续型数据离散化——等宽分组&等频分组

import pandas as pd
import numpy as np

# 加载数据
detail = pd.read_excel("../day05/meal_order_detail.xlsx")
# print("detail ：", detail)
print("detail 的列索引：", detail.columns)

# 将dishes_name 转化为数值型数据
# 哑变量矩阵转化
res = pd.get_dummies(
    data=detail.loc[:,"dishes_name"],
    prefix_sep="_",
    prefix="菜品"
)
print("转化之后的结果res:
",res)
# res.to_csv("./hh.csv")


# 身高 150 - 190  每位同学 都是一个具体的身高---连续的小数
# 将连续型数据转化为类别数据 ----离散化
# 分组
print("菜品单价的最大值与最小值：", detail.loc[:, "amounts"].max(), detail.loc[:, "amounts"].min())
# 将detail 里面的amounts 数据进行离散化
# detail.loc[:, "amounts"] = pd.cut(detail.loc[:, "amounts"], bins=5)

# 自定义分组
# # 等宽分组
# # 1、指定分组个数
group_num = 5
# # # 2、计算最大值与最小值的极差
ptp = detail.loc[:, "amounts"].max() - detail.loc[:, "amounts"].min()
# # # 3、确定步长
step = int(np.ceil(ptp / group_num))
# # # 4、确定分组的区间的节点
bins = np.arange(detail.loc[:, "amounts"].min(), detail.loc[:, "amounts"].max() + step, step)
print(bins)
# # 5、指定自定义分组
# # include_lowest ---指定包含最小值
detail.loc[:, "amounts"] = pd.cut(detail.loc[:, "amounts"], bins=bins, include_lowest=True)

# 等频分组
# 1、计算分位数
# bins = detail.loc[:, "amounts"].quantile(q=np.arange(0, 1 + 1 / 5, 1 / 5))
# print(bins)
# # include_lowest ---指定包含最小值
# detail.loc[:, "amounts"] = pd.cut(detail.loc[:, "amounts"], bins=bins, include_lowest=True)
#
# print(detail.loc[:, "amounts"])
# #
# # 统计每一个组内的个数
# res_counts = pd.value_counts(detail.loc[:, "amounts"])
# print("res_counts:
", res_counts)