利用python 封装函数，计算加法，乘法， 除法贡献度

已于 2022-08-26 17:12:59 修改
阅读量735 收藏 1
点赞数 2
文章标签： python 机器学习 numpy
于 2022-08-26 17:09:06 首次发布
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/weixin_46147435/article/details/126547086
版权

上篇文章简单介绍了一下加法，乘法，除法贡献度的计算方法，并手动实现了一下过程，文章链接：加法、乘法、除法贡献度计算实现

这篇文章想要把贡献度计算方法封装成一个函数，在传入json输入的时候，自动计算输出结果，其中乘法贡献度的计算方法有做更改，具体如下

一、加载需要的包

import pandas as pd
import numpy as np
import json

二、读取数据（加法，其中A001+A002+A003 = A01）

df_add = pd.read_excel("C:/Users/supaur/Desktop/value.xlsx")

df_add
index_numindex_nametimeperiodvalue
01A012022-08-22day1000
12A0012022-08-22day300
23A0022022-08-22day200
34A0032022-08-22day500
41A012022-08-21day1900
52A0012022-08-21day1000
63A0022022-08-21day500
74A0032022-08-21day400

三、查看传入的加法json并解析出相应的参数

json_file = "C:/Users/supaur/Desktop/add.json"
with open(json_file, "r") as fp:
    data = json.load(fp)
print(data)
date = data["date"]
target1 = data["target1"]
target2 = data["target2"]
target2_1 = data["target2_1"]
target2_2 = data["target2_2"]
type = data["type"]
timetype = data["timetype"]

{'date': ['20220821', '20220822'], 'target1': 'A01', 'target2': ['A001', 'A002', 'A003'], 'target2_1': [], 'target2_2': [], 'type': 'add', 'timetype': 'day'}

四、编写加法贡献度函数

# 生成两个df，分别为参数date里的两个日期，打印出来看看结果
df1_add = df_add[(df_add["time"] == date[0]) & (df_add["period"] == timetype)]
df2_add = df_add[(df_add["time"] == date[1]) & (df_add["period"] == timetype)]
print(df1_add)
print(df2_add)

   index_num index_name       time period  value
4          1        A01 2022-08-21    day   1900
5          2       A001 2022-08-21    day   1000
6          3       A002 2022-08-21    day    500
7          4       A003 2022-08-21    day    400
   index_num index_name       time period  value
0          1        A01 2022-08-22    day   1000
1          2       A001 2022-08-22    day    300
2          3       A002 2022-08-22    day    200
3          4       A003 2022-08-22    day    500

# df1与df2 按照字段名称拼接成下面格式
df3_add = pd.merge(df1_add, df2_add, how="inner", left_on=["index_name"], right_on=["index_name"],suffixes=("_1", "_2"))
df3_add
index_num_1index_nametime_1period_1value_1index_num_2time_2period_2value_2
01A012022-08-21day190012022-08-22day1000
12A0012022-08-21day100022022-08-22day300
23A0022022-08-21day50032022-08-22day200
34A0032022-08-21day40042022-08-22day500
# 添加一列字段，得到对应字段两个日期的差值
df3_add["D_value"] = df3_add.loc[:, "value_2"] - df3_add.loc[:, "value_1"]
df3_add
index_num_1index_nametime_1period_1value_1index_num_2time_2period_2value_2D_value
01A012022-08-21day190012022-08-22day1000-900
12A0012022-08-21day100022022-08-22day300-700
23A0022022-08-21day50032022-08-22day200-300
34A0032022-08-21day40042022-08-22day500100
# 利用for循环，计算每个target2的贡献度，其中计算方法为target2所在的行，D_value所在列 /target1所在的行，D_value所在列，得到每个target2的贡献度
for i in target2:
    df3_add["%s_contribution" % i] = round(df3_add.loc[df3_add["index_name"] == i, "D_value"].tolist()[0] / df3_add.loc[df3_add["index_name"] == target1, "D_value"].tolist()[0] *100,2)
df3_add
index_num_1index_nametime_1period_1value_1index_num_2time_2period_2value_2D_valueA001_contributionA002_contributionA003_contribution
01A012022-08-21day190012022-08-22day1000-90077.7833.33-11.11
12A0012022-08-21day100022022-08-22day300-70077.7833.33-11.11
23A0022022-08-21day50032022-08-22day200-30077.7833.33-11.11
34A0032022-08-21day40042022-08-22day50010077.7833.33-11.11
# 把以上代码组合封装成functionadd函数，创建一个空列表，把贡献度值添加到列表里，并调用该函数打印结果如下：
def functionadd(date,target1,target2,timetype):

    df1_add = df_add[(df_add["time"] == date[0]) & (df_add["period"] == timetype)] 
    df2_add = df_add[(df_add["time"] == date[1]) & (df_add["period"] == timetype)]
    df3_add = pd.merge(df1, df2, how="inner", left_on=["index_name"], right_on=["index_name"],suffixes=("_1", "_2"))
    df3_add["D_value"] = df3_add.loc[:, "value_2"] - df3_add.loc[:, "value_1"]
    re = []
    for i in target2:
        df3_add["%s_contribution" % i] = round(df3_add.loc[df3_add["index_name"] == i, "D_value"].tolist()[0] / df3_add.loc[df3_add["index_name"] == target1, "D_value"].tolist()[0] *100,2)
        re.append(("%s_contribution" % i, df3_add["%s_contribution" % i].tolist()[0]))
    return re
        
print(functionadd(date,target1,target2,timetype))

[('A001_contribution', 29.94), ('A002_contribution', 19.97), ('A003_contribution', 49.99)]

五、乘法跟加法类似，再开始加法之前先对各指标求log，变成加法计算

# 查看传入的乘法json并解析出相应的参数
json_file = "C:/Users/supaur/Desktop/times.json"
with open(json_file, "r") as fp:
    data = json.load(fp)
print(data)
date = data["date"]
target1 = data["target1"]
target2 = data["target2"]
target2_1 = data["target2_1"]
target2_2 = data["target2_2"]
type = data["type"]
timetype = data["timetype"]

{'date': ['20220801', '20220901'], 'target1': 'B01', 'target2': ['B001', 'B002'], 'target2_1': [], 'target2_2': [], 'type': 'times', 'timetype': 'month'}

# 查看乘法数据，其中A001*A002*A003 = A01 , B001*B002 = B01
df_times = pd.read_excel("C:/Users/supaur/Desktop/times.xlsx") 
df_times
index_numindex_nametimeperiodvalue
01A012022-08-22day30000.00
12A0012022-08-22day0.30
23A0022022-08-22day200.00
34A0032022-08-22day500.00
41A012022-08-21day20000.00
52A0012022-08-21day0.10
63A0022022-08-21day500.00
74A0032022-08-21day400.00
81A012022-08-20day32000.00
92A0012022-08-20day0.20
103A0022022-08-20day400.00
114A0032022-08-20day400.00
125B012022-08-01month150000.00
136B0012022-08-01month0.30
147B0022022-08-01month500000.00
155B012022-09-01month107500.00
166B0012022-09-01month0.25
177B0022022-09-01month430000.00
185B012022-10-01month185500.00
196B0012022-10-01month0.35
207B0022022-10-01month530000.00
# 编写乘法函数，加法是直接求两日期差值，乘法是先取对数log，再变成加法计算
def functiontimes(date,target1,target2,timetype):

    df1_times = df_times[(df_times["time"] == date[0]) & (df_times["period"] == timetype)]
    df2_times = df_times[(df_times["time"] == date[1]) & (df_times["period"] == timetype)]
    df3_times = pd.merge(df1_times, df2_times, how="inner", left_on=["index_name"], right_on=["index_name"],suffixes=("_1", "_2"))
    df3_times["D_value"] = np.log10(df3_times.loc[:, "value_2"]) - np.log10(df3_times.loc[:, "value_1"])
    re = []
    for i in target2:
        df3_times["%s_contribution" % i] = round(df3_times.loc[df3_times["index_name"] == i, "D_value"].tolist()[0] / df3_times.loc[df3_times["index_name"] == target1, "D_value"].tolist()[0] *100,2)
        re.append(("%s_contribution" % i, df3_times["%s_contribution" % i].tolist()[0]))
    return re


print(functiontimes(date,target1,target2,timetype))

[('B001_contribution', 54.73), ('B002_contribution', 45.27)]

六、除法贡献度函数

# 查看传入的除法json并解析出相应的参数，除法要比加法和乘法多两个参数target2_1，target2_2
json_file = "C:/Users/supaur/Desktop/div.json"
with open(json_file, "r") as fp:
    data = json.load(fp)
print(data)
date = data["date"]
target1 = data["target1"]
target2 = data["target2"]
target2_1 = data["target2_1"]
target2_2 = data["target2_2"]
type = data["type"]
timetype = data["timetype"]

{'date': ['20220822', '20220823'], 'target1': 'A01', 'target2': ['A001', 'A002', 'A003'], 'target2_1': ['A001_1', 'A002_1', 'A003_1'], 'target2_2': ['A001_2', 'A002_2', 'A003_2'], 'type': 'div', 'timetype': 'day'}

# 查看除法数据，其中A001 = A001_1/A001_2 , A002 = A002_1/A002_2 , 以此类推，A01 = （A001_1 + A002_1 + A003_1) / (A001_2 + A002_2 + A003_2)
df_div = pd.read_excel("C:/Users/supaur/Desktop/div.xlsx")
df_div
index_numindex_nametimeperiodvalue
01A012022-08-22day0.840546
12A0012022-08-22day0.882581
22A001_12022-08-22day684.000000
32A001_22022-08-22day775.000000
43A0022022-08-22day0.732308
53A002_12022-08-22day714.000000
63A002_22022-08-22day975.000000
74A0032022-08-22day0.930061
84A003_12022-08-22day758.000000
94A003_22022-08-22day815.000000
101A012022-08-23day0.448011
112A0012022-08-23day0.696589
122A001_12022-08-23day388.000000
132A001_22022-08-23day557.000000
143A0022022-08-23day0.415882
153A002_12022-08-23day309.000000
163A002_22022-08-23day743.000000
174A0032022-08-23day0.322368
184A003_12022-08-23day294.000000
194A003_22022-08-23day912.000000
# 与加法和乘法一样 生成两个df，分别为参数date里的两个日期
df1_div = df_div[(df_div["time"] == date[0]) & (df_div["period"] == timetype)]
df2_div = df_div[(df_div["time"] == date[1]) & (df_div["period"] == timetype)]
print(df1_div)
print(df2_div)

   index_num index_name       time period       value
0          1        A01 2022-08-22    day    0.840546
1          2       A001 2022-08-22    day    0.882581
2          2     A001_1 2022-08-22    day  684.000000
3          2     A001_2 2022-08-22    day  775.000000
4          3       A002 2022-08-22    day    0.732308
5          3     A002_1 2022-08-22    day  714.000000
6          3     A002_2 2022-08-22    day  975.000000
7          4       A003 2022-08-22    day    0.930061
8          4     A003_1 2022-08-22    day  758.000000
9          4     A003_2 2022-08-22    day  815.000000
    index_num index_name       time period       value
10          1        A01 2022-08-23    day    0.448011
11          2       A001 2022-08-23    day    0.696589
12          2     A001_1 2022-08-23    day  388.000000
13          2     A001_2 2022-08-23    day  557.000000
14          3       A002 2022-08-23    day    0.415882
15          3     A002_1 2022-08-23    day  309.000000
16          3     A002_2 2022-08-23    day  743.000000
17          4       A003 2022-08-23    day    0.322368
18          4     A003_1 2022-08-23    day  294.000000
19          4     A003_2 2022-08-23    day  912.000000

# 求所有target2_1 的和，新增列为sum_1
sum_1 = 0
for i in target2_1:
    sum_1 = sum_1 + df1_div.loc[df1_div["index_name"] == i , "value"].tolist()[0]
df1_div["sum_1"] = sum_1
df1_div
index_numindex_nametimeperiodvaluesum_1
01A012022-08-22day0.8405462156.0
12A0012022-08-22day0.8825812156.0
22A001_12022-08-22day684.0000002156.0
32A001_22022-08-22day775.0000002156.0
43A0022022-08-22day0.7323082156.0
53A002_12022-08-22day714.0000002156.0
63A002_22022-08-22day975.0000002156.0
74A0032022-08-22day0.9300612156.0
84A003_12022-08-22day758.0000002156.0
94A003_22022-08-22day815.0000002156.0
# 求所有target2_2 的和，新增列为sum_2
sum_2 = 0
for i in target2_2:
    sum_2 = sum_2 + df1_div.loc[df1_div["index_name"] == i , "value"].tolist()[0]
df1_div["sum_2"] = sum_2
df1_div
index_numindex_nametimeperiodvaluesum_1sum_2
01A012022-08-22day0.8405462156.02565.0
12A0012022-08-22day0.8825812156.02565.0
22A001_12022-08-22day684.0000002156.02565.0
32A001_22022-08-22day775.0000002156.02565.0
43A0022022-08-22day0.7323082156.02565.0
53A002_12022-08-22day714.0000002156.02565.0
63A002_22022-08-22day975.0000002156.02565.0
74A0032022-08-22day0.9300612156.02565.0
84A003_12022-08-22day758.0000002156.02565.0
94A003_22022-08-22day815.0000002156.02565.0
# df1_div, df2_div 按照字段名称拼接成下面格式
df3_div = pd.merge(df1_div, df2_div, how="inner", left_on=["index_name"], right_on=["index_name"],suffixes=("_1", "_2"))
df3_div
index_num_1index_nametime_1period_1value_1sum_1sum_2index_num_2time_2period_2value_2
01A012022-08-22day0.8405462156.02565.012022-08-23day0.448011
12A0012022-08-22day0.8825812156.02565.022022-08-23day0.696589
22A001_12022-08-22day684.0000002156.02565.022022-08-23day388.000000
32A001_22022-08-22day775.0000002156.02565.022022-08-23day557.000000
43A0022022-08-22day0.7323082156.02565.032022-08-23day0.415882
53A002_12022-08-22day714.0000002156.02565.032022-08-23day309.000000
63A002_22022-08-22day975.0000002156.02565.032022-08-23day743.000000
74A0032022-08-22day0.9300612156.02565.042022-08-23day0.322368
84A003_12022-08-22day758.0000002156.02565.042022-08-23day294.000000
94A003_22022-08-22day815.0000002156.02565.042022-08-23day912.000000
# 添加一列字段，得到对应字段两个日期的差值
df3_div["D_value"] = df3_div.loc[:, "value_2"] - df3_div.loc[:, "value_1"]
df3_div
index_num_1index_nametime_1period_1value_1sum_1sum_2index_num_2time_2period_2value_2D_value
01A012022-08-22day0.8405462156.02565.012022-08-23day0.448011-0.392535
12A0012022-08-22day0.8825812156.02565.022022-08-23day0.696589-0.185992
22A001_12022-08-22day684.0000002156.02565.022022-08-23day388.000000-296.000000
32A001_22022-08-22day775.0000002156.02565.022022-08-23day557.000000-218.000000
43A0022022-08-22day0.7323082156.02565.032022-08-23day0.415882-0.316426
53A002_12022-08-22day714.0000002156.02565.032022-08-23day309.000000-405.000000
63A002_22022-08-22day975.0000002156.02565.032022-08-23day743.000000-232.000000
74A0032022-08-22day0.9300612156.02565.042022-08-23day0.322368-0.607693
84A003_12022-08-22day758.0000002156.02565.042022-08-23day294.000000-464.000000
94A003_22022-08-22day815.0000002156.02565.042022-08-23day912.00000097.000000
# 求每个target2 的贡献度，计算方法为每个 target2_1 的 D_value + sum_1 的和 / 每个 target2_2 的 D_value + sum_2 的和  - target1 的 value_1
# 再对target2 的贡献度求和
sum = 0
for (a ,i ,j) in zip(target2,target2_1,target2_2):
    df3_div["%s_contribution" % a] = (df3_div.loc[df3_div["index_name"] == i, "sum_1"].tolist()[0] + df3_div.loc[df3_div["index_name"] == i,"D_value"].tolist()[0]) / (df3_div.loc[df3_div["index_name"] == j, "sum_2"].tolist()[0] + df3_div.loc[df3_div["index_name"] == j,"D_value"].tolist()[0]) - df3_div.loc[df3_div["index_name"] == target1, "value_1"].tolist()[0]
    sum = sum + df3_div["%s_contribution" % a]
df3_div["sum"] = sum
df3_div
index_num_1index_nametime_1period_1value_1sum_1sum_2index_num_2time_2period_2value_2D_valueA001_contributionA002_contributionA003_contributionsum
01A012022-08-22day0.8405462156.02565.012022-08-23day0.448011-0.392535-0.048045-0.09001-0.204933-0.342988
12A0012022-08-22day0.8825812156.02565.022022-08-23day0.696589-0.185992-0.048045-0.09001-0.204933-0.342988
22A001_12022-08-22day684.0000002156.02565.022022-08-23day388.000000-296.000000-0.048045-0.09001-0.204933-0.342988
32A001_22022-08-22day775.0000002156.02565.022022-08-23day557.000000-218.000000-0.048045-0.09001-0.204933-0.342988
43A0022022-08-22day0.7323082156.02565.032022-08-23day0.415882-0.316426-0.048045-0.09001-0.204933-0.342988
53A002_12022-08-22day714.0000002156.02565.032022-08-23day309.000000-405.000000-0.048045-0.09001-0.204933-0.342988
63A002_22022-08-22day975.0000002156.02565.032022-08-23day743.000000-232.000000-0.048045-0.09001-0.204933-0.342988
74A0032022-08-22day0.9300612156.02565.042022-08-23day0.322368-0.607693-0.048045-0.09001-0.204933-0.342988
84A003_12022-08-22day758.0000002156.02565.042022-08-23day294.000000-464.000000-0.048045-0.09001-0.204933-0.342988
94A003_22022-08-22day815.0000002156.02565.042022-08-23day912.00000097.000000-0.048045-0.09001-0.204933-0.342988
# 标准化，每个target2贡献度值/ 所有target2 总和 得到每个 target2 标准化后的贡献度
for b in target2:
    df3_div["%s_nor_contribution" % b] = round(df3_div.loc[df3_div["index_name"] == b, "%s_contribution" % b].tolist()[0] / df3_div.loc[df3_div["index_name"] == b, "sum"].tolist()[0] * 100, 2)
df3_div
index_num_1index_nametime_1period_1value_1sum_1sum_2index_num_2time_2period_2value_2D_valueA001_contributionA002_contributionA003_contributionsumA001_nor_contributionA002_nor_contributionA003_nor_contribution
01A012022-08-22day0.8405462156.02565.012022-08-23day0.448011-0.392535-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
12A0012022-08-22day0.8825812156.02565.022022-08-23day0.696589-0.185992-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
22A001_12022-08-22day684.0000002156.02565.022022-08-23day388.000000-296.000000-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
32A001_22022-08-22day775.0000002156.02565.022022-08-23day557.000000-218.000000-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
43A0022022-08-22day0.7323082156.02565.032022-08-23day0.415882-0.316426-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
53A002_12022-08-22day714.0000002156.02565.032022-08-23day309.000000-405.000000-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
63A002_22022-08-22day975.0000002156.02565.032022-08-23day743.000000-232.000000-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
74A0032022-08-22day0.9300612156.02565.042022-08-23day0.322368-0.607693-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
84A003_12022-08-22day758.0000002156.02565.042022-08-23day294.000000-464.000000-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
94A003_22022-08-22day815.0000002156.02565.042022-08-23day912.00000097.000000-0.048045-0.09001-0.204933-0.34298814.0126.2459.75
# 对上述代码封装成除法贡献度函数
def functiondiv(date,target1,target2,target2_1,target2_2,timetype):

    df1_div = df_div[(df_div["time"] == date[0]) & (df_div["period"] == timetype)]
    df2_div = df_div[(df_div["time"] == date[1]) & (df_div["period"] == timetype)]
    sum_1 = 0
    for i in target2_1:
        sum_1 = sum_1 + df1_div.loc[df1_div["index_name"] == i , "value"].tolist()[0]
    df1_div["sum_1"] = sum_1
    sum_2 = 0
    for i in target2_2:
        sum_2 = sum_2 + df1_div.loc[df1_div["index_name"] == i , "value"].tolist()[0]
    df1_div["sum_2"] = sum_2
    df3_div = pd.merge(df1_div, df2_div, how="inner", left_on=["index_name"], right_on=["index_name"],suffixes=("_1", "_2"))
    df4_div = df3_div.loc[:, ["index_num_1", "index_name", "time_1", "time_2", "period_1", "value_1", "value_2","sum_1","sum_2"]]
    df4_div["D_value"] = df4_div.loc[:, "value_2"] - df4_div.loc[:, "value_1"]
    sum = 0
    for (a ,i ,j) in zip(target2,target2_1,target2_2):
        df4_div["%s_contribution" % a] = (df4_div.loc[df4_div["index_name"] == i, "sum_1"].tolist()[0] + df4_div.loc[df4_div["index_name"] == i,"D_value"].tolist()[0]) / (df4_div.loc[df4_div["index_name"] == j, "sum_2"].tolist()[0] + df4_div.loc[df4_div["index_name"] == j,"D_value"].tolist()[0]) - df4_div.loc[df4_div["index_name"] == target1, "value_1"].tolist()[0]
        sum = sum + df4_div["%s_contribution" % a]
    df4_div["sum"] = sum
    re = []
    for b in target2:
        df4_div["%s_nor_contribution" % b] = round(df4_div.loc[df4_div["index_name"] == b, "%s_contribution" % b].tolist()[0] / df4_div.loc[df4_div["index_name"] == b, "sum"].tolist()[0] * 100, 2)
        re.append(("%s_nor_contribution" % b, df4_div["%s_nor_contribution" % b].tolist()[0]))
    return re

print(functiondiv(date, target1, target2,target2_1,target2_2, timetype))

[('A001_nor_contribution', 14.01), ('A002_nor_contribution', 26.24), ('A003_nor_contribution', 59.75)]
weixin_46147435
关注
  • 2
    点赞
  • 踩
  • 1
    收藏
    觉得还不错? 一键收藏
  • 3
    评论
如何计算指标波动贡献率？（附Pandas实现）
SeizeeveryDay的博客
01-08 1190
大家好，我是阿粥“为什么这个月销售额提升了30%？”“为什么转化率又降了，同比竟然降低了42%，什么原因导致的呢？”这些都是数据分析师在工作中经常会遇到的问题，甚至有些基础岗的数据分析师要花80%以上的精力处理这类问题：指标降低或者提升了多少，以及波动的原因是什么。这类问题可以概括为指标波动归因分析，很多时候用的是根据经验探索拆分的办法，非常占用数据分析师的时间，而且数据分析师获得的价值感也不高。...
除法类型复合指标异动贡献度计算
fff2zrx的博客
09-21 729
除法类型复合指标异动贡献度的python计算模版
归因分析中各种指标因子贡献度计算方法总结
最新发布
baidu_26137595的博客
03-29 5510
在归因分析计算因子贡献度时，加法指标、乘法指标、除法指标由于指标类型的不同，所用的方法也不同，本文总结三种不同类型指标计算因子贡献度的方法，并进行具体的案例分析、公式推导、优缺点分析。
数据探索（数据特征分析）④—Python分布分析、对比分析、统计量分析、期性分析、贡献度分析、相关性分析
qq_37746855的博客
12-09 7296
数据探索（数据特征分析）—Python数据特征分析分布分析定量数据的分布分析遵循的主要原则有实例Python实现代码实现定性数据的分布分析对比分析对比分析主要有以下两种形式：统计量分析周期性分析贡献度分析相关性分析 数据特征分析 对数据进行质量分析以后，接下来就是对数据做特征分析。一般可通过绘制图表、计算某些特征量等手段进行数据的特征分析。 分布分析 分布分析能揭示数据的分布特征和分布类型，便于发现某些特大或特小的可疑值。对于定量数据，欲了解其分布形式，是对称的、还是非对称的，可做出频率分布表、绘制频率分布
高精度加法、减法、乘法的函数封装
m0_73020012的博客
08-16 309
在c++的运算中，我们有的时候会用到非常大的数据，以至于__int128也无法存下，这个时候，我们就需要使用高精度数进行运算。但是，高精度数的写法非常的不方便，所以，本文主要讲解如何把高精度数封装为结构体类型，如何把高精度数的计算封装为通用函数。
加法、乘法、除法贡献度计算实现
weixin_46147435的博客
08-17 2116
加法、乘法、除法贡献度计算实现
Python3 根据乘法口诀生成除法口算题
03-09
Python3 根据乘法口诀生成除法口算题。用到的简单的技术点：当前时间，读取文件、写入文件。使用Python。希望大家多多交流。QQ：2444756311.注明：Python
Python实现矩阵加法和乘法的方法分析
01-02
本文实例讲述了Python实现矩阵加法和乘法的方法。分享给大家供大家参考，具体如下： 本来以为python的矩阵用list表示出来应该很简单可以搞。。其实发现有大学问。 这里贴出我写的特别不pythonic的矩阵加法，作为反例...
python封装的异常值处理函数（包括箱线图去除异常值等）
12-21
本篇内容主要介绍了如何封装一个 Python 函数 `outliers_proc`，该函数用于处理数据集中的异常值，特别是通过箱线图法去除异常值。 首先，我们来看函数 `outliers_proc`，它接受三个参数：`data`（Pandas DataFrame...
python微元法计算函数曲线长度的方法
12-23
，令积分函数 f(x,y,z) 为1，即计算曲线的长度，将其微元化： 其中 根据此时便可在python编程实现，给出4个例子，代码中已有详细注释，不再赘述 ''' 计算曲线长度，根据线积分公式： \int_A^Bf(x,y,z)dl，令积分...
异动分析（四）利用Python计算指标贡献度
微信知乎搜索HsuHeinrich
03-02 1万+
利用Python计算指标贡献度
Python 对加减乘除分别封装一个函数进行计算，参数是两个数，返回值是计算结果
weixin_42218889的博客
08-07 8909
1.对加减乘除分别封装一个函数进行计算，参数是两个数，返回值是计算结果 def add(x, y): """ :param x: :param y: :return:返回结果 """ print(x + y) return add(3,4) def sub(x, y): """ :param x: :param ...
封装加减乘除方法
irving2372357的博客
12-10 1090
public class Operation1 { // 加法 public int add(int x,int y) { return x+y; } //减法 public int sub(int x,int y) { return x-y; } //乘法 public int mul(int x, int y) { return xy; } //除法 public int div(int x,...
指标异动的贡献度量化归因
weixin_45520028的博客
03-02 2893
目录背景目标方法思路子维度异动的贡献度量化指标类型一：加法指标类型二：乘法指标类型三：除法指标类型四：复合指标指标类型五：其他指标诊断展示结论 背景 数据监控，核心聚焦在子维度粒度数据波动的解读，但是不便于定位异常波动水平，以及进行指标下钻分析，难以直接定位指标变动来源。难以量化子维度贡献度 例如：11点准时率下滑，但通过分省份数据很难直接各省份波动程度，也不便于定位各省份对11点准时率下滑影响。 目标 自动化实现指标异动监控预警及变化来源量化定位。 方法思路 确定各个维度层级异常预警逻辑，简单定位异常省份
Python数据分析工具，主要有哪些？
多智时代的博客
04-02 3661
python数据分析工具一：IPython IPython是一个在多种编程语言之间进行交互计算的命令行shell，最开始是用python开发的，提供增强的内省，富媒体，扩展的shell语法，tab补全，丰富的历史等功能。IPython提供了如下特性： 更强的交互shell(基于Qt的终端) 一个基于浏览器的记事本，支持代码，纯文本，数学公式，内置图表和其他富媒体 支持交互数据可视化和图形界面工具 ...
归因分析计算因子贡献度常见的方法
热门推荐
Hubans的博客
01-31 2万+
在归因分析中，我们一般都需要计算出每个因子的贡献度是多少，比如产品DAU上升，对年龄段维度进行拆解，看是不同年龄段的用户对DAU上升的贡献度是多少，一般根据指标的类型，计算贡献度的方法也不一样，下面就列出一些常见的归因分析贡献度的计算方法。
【阿里妈妈营销科学系列】第八篇：增强分析在营销分析场景下的实现和应用...
阿里妈妈技术
02-23 2053
丨目录：· 背景· 业务场景· 基本概念· 阿里妈妈智能洞察系统Autolnsight· 应用场景和案例· 总结· 关于我们▐ 背景随着以数据驱动业务增长的发展趋势，“数据驱动力”在精益运...
js函数封装数字相加的几种方法
屈小康
03-30 824
1.通过内置方法 arguments function sum(){ let count = 0 for(let i = 0 ; i < arguments.length ; i++){ if(isNaN(arguments[i])){ } else{ arguments[i] = Number(arguments[i]) count = arguments[i]+ count }n } return count
加、减、乘、除 函数封装
这是我的博客
07-17 1160
经常在进行数据的计算时，出现数据的小数后出现好多位或者数值类型不一致导致计算有误 加： function floatAdd(arg1, arg2) { var r1, r2, m; try { r1 = arg1.toString().split(".")[1].length } catch (e) { r1 = 0 } try { r2 = arg2.toString().s...
利用python计算多项式乘法
05-16
可以使用Python中的NumPy库来计算多项式乘法。以下是一个示例代码： ```python import numpy as np def poly_multiply(a, b): m, n = len(a), len(b) res = np.zeros(m + n - 1) for i in range(m): for j in ...

“相关推荐”对你有帮助么？

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论 3
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值