Question

我的目标是获得一些季节性销售产品的预测销售数据。

我已经设法通过forecast包和arima方法获得了针对单个产品的预测。现在，我可以访问历史销售数据，其中包含不同地区每月销售的商品数量。历史数据可以追溯到几年前。为了概述数据的组织方式，我在下面编写了一段摘录：

History <- read.table(sep = ";",text = "
Region;Part.number;Qty;Period;Year;Month;Plant;Material.ID;Subgroup.ID;Group.ID;Product.Subgroup;Product.Group;Not.Consider
AA;1.234;5;01/01/2008;2008;1;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;8;01/01/2008;2008;1;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;16;01/01/2008;2008;1;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;5;01/01/2008;2008;1;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;14;01/01/2008;2008;1;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;3;01/01/2008;2008;1;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;9;01/01/2008;2008;1;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;4;01/01/2008;2008;1;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;6;01/01/2008;2008;1;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;3;01/01/2008;2008;1;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;27;01/02/2008;2008;2;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;42;01/02/2008;2008;2;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;80;01/02/2008;2008;2;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;27;01/02/2008;2008;2;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;72;01/02/2008;2008;2;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;18;01/02/2008;2008;2;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;46;01/02/2008;2008;2;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;20;01/02/2008;2008;2;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;33;01/02/2008;2008;2;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;16;01/02/2008;2008;2;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;81;01/03/2008;2008;3;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;126;01/03/2008;2008;3;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;244;01/03/2008;2008;3;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;82;01/03/2008;2008;3;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;219;01/03/2008;2008;3;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;56;01/03/2008;2008;3;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;138;01/03/2008;2008;3;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;60;01/03/2008;2008;3;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;101;01/03/2008;2008;3;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;52;01/03/2008;2008;3;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;165;01/04/2008;2008;4;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;255;01/04/2008;2008;4;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;490;01/04/2008;2008;4;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;167;01/04/2008;2008;4;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;440;01/04/2008;2008;4;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;115;01/04/2008;2008;4;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;277;01/04/2008;2008;4;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;120;01/04/2008;2008;4;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;204;01/04/2008;2008;4;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;106;01/04/2008;2008;4;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;234;01/05/2008;2008;5;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;357;01/05/2008;2008;5;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;687;01/05/2008;2008;5;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;235;01/05/2008;2008;5;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;619;01/05/2008;2008;5;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;161;01/05/2008;2008;5;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;389;01/05/2008;2008;5;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;172;01/05/2008;2008;5;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;289;01/05/2008;2008;5;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;150;01/05/2008;2008;5;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;237;01/06/2008;2008;6;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;361;01/06/2008;2008;6;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;690;01/06/2008;2008;6;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;237;01/06/2008;2008;6;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;620;01/06/2008;2008;6;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;164;01/06/2008;2008;6;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;393;01/06/2008;2008;6;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;174;01/06/2008;2008;6;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;293;01/06/2008;2008;6;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;153;01/06/2008;2008;6;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;173;01/07/2008;2008;7;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;260;01/07/2008;2008;7;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;496;01/07/2008;2008;7;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;172;01/07/2008;2008;7;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;443;01/07/2008;2008;7;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;119;01/07/2008;2008;7;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;281;01/07/2008;2008;7;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;125;01/07/2008;2008;7;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;212;01/07/2008;2008;7;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;112;01/07/2008;2008;7;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;88;01/08/2008;2008;8;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;131;01/08/2008;2008;8;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;252;01/08/2008;2008;8;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;88;01/08/2008;2008;8;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;225;01/08/2008;2008;8;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;63;01/08/2008;2008;8;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;141;01/08/2008;2008;8;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;63;01/08/2008;2008;8;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;106;01/08/2008;2008;8;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;58;01/08/2008;2008;8;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;31;01/09/2008;2008;9;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;47;01/09/2008;2008;9;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;88;01/09/2008;2008;9;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;31;01/09/2008;2008;9;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;79;01/09/2008;2008;9;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;24;01/09/2008;2008;9;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;48;01/09/2008;2008;9;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;24;01/09/2008;2008;9;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;38;01/09/2008;2008;9;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;23;01/09/2008;2008;9;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;7;01/10/2008;2008;10;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;11;01/10/2008;2008;10;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;20;01/10/2008;2008;10;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;8;01/10/2008;2008;10;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;17;01/10/2008;2008;10;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;5;01/10/2008;2008;10;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;11;01/10/2008;2008;10;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;6;01/10/2008;2008;10;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;10;01/10/2008;2008;10;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;5;01/10/2008;2008;10;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;2;01/11/2008;2008;11;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;2;01/11/2008;2008;11;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;3;01/11/2008;2008;11;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;2;01/11/2008;2008;11;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;3;01/11/2008;2008;11;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;4;01/11/2008;2008;11;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;3;01/11/2008;2008;11;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;2;01/11/2008;2008;11;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;5;01/11/2008;2008;11;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;3;01/11/2008;2008;11;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0
AA;1.234;1;01/12/2008;2008;12;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;0;01/12/2008;2008;12;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;1.234;2;01/12/2008;2008;12;555;1.234_AA_555;Product A_AA_555;Group1_AA_555;Product A;Group1;0
AA;2.345;1;01/12/2008;2008;12;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
AA;2.345;2;01/12/2008;2008;12;555;2.345_AA_555;Product B_AA_555;Group1_AA_555;Product B;Group1;0
BB;2.345;3;01/12/2008;2008;12;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;3;01/12/2008;2008;12;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;2.345;3;01/12/2008;2008;12;444;2.345_BB_444;Product B_BB_444;Group1_BB_444;Product B;Group1;0
BB;3.456;3;01/12/2008;2008;12;444;3.456_BB_444;Product C_BB_444;Group2_BB_444;Product C;Group2;0
BB;3.456;4;01/12/2008;2008;12;00X;3.456_BB_00X;Product C_BB_00X;Group2_BB_00X;Product C;Group2;0

",
header=TRUE)

这是头表的显示方式：

  Region Part.number Qty     Period Year Month Plant  Material.ID      Subgroup.ID      Group.ID Product.Subgroup Product.Group Not.Consider
1     AA       1.234   5 01/01/2008 2008     1   555 1.234_AA_555 Product A_AA_555 Group1_AA_555        Product A        Group1            0
2     AA       1.234   8 01/01/2008 2008     1   555 1.234_AA_555 Product A_AA_555 Group1_AA_555        Product A        Group1            0
3     AA       1.234  16 01/01/2008 2008     1   555 1.234_AA_555 Product A_AA_555 Group1_AA_555        Product A        Group1            0
4     AA       2.345   5 01/01/2008 2008     1   555 2.345_AA_555 Product B_AA_555 Group1_AA_555        Product B        Group1            0
5     AA       2.345  14 01/01/2008 2008     1   555 2.345_AA_555 Product B_AA_555 Group1_AA_555        Product B        Group1            0
6     BB       2.345   3 01/01/2008 2008     1   444 2.345_BB_444 Product B_BB_444 Group1_BB_444        Product B        Group1            0

中间目标是获得一个包含2015年预测值的表格：

Region  Part.number  January  February  March  April  May  ...  Total
AA            1.234        2         4      9      9    5  ...     29

我正在努力解决的问题是如何将数据提供给forecast::stl，以及如何为所有不同的Region＆amp; Part.number组合提供数据。

欣赏提示。

Answer 1

# load some useful libraries
library(forecast)
library(dplyr)
library(stringr)
library(lubridate)

# reformat Period column to Date variable
df <-
  History %>% 
  mutate(Period = as.Date(dmy(Period)))

# prepeare a vector of possible Region-Part.number combinations
filtered <- 
  df %>% 
  select(Region, Part.number) %>% 
  unique(.)

# filter your data using above defined object called "filtered"
my_df <- 
  df %>% 
  filter(Region == filtered[1, 1],
         Part.number == filtered[1, 2])

# select only relevant columns, prepeare name for a time series
my_ts_df <-
  my_df %>% 
  select(Region, Part.number, Qty, Period) %>% 
  mutate(reg_part = str_c(Region, Part.number)) %>% 
  select(reg_part, Period, Qty)

my_ts_name <- 
  my_ts_df %>% 
  select(reg_part) %>% 
  unique(.) %>% 
  unlist(.) %>% 
  unname(.)

使用如下的ts（）命令：

ts(my_ts_df$Qty, start = first(my_ts_df$Period), end = last(my_ts_df$Period), freq = 12)

你应该能够构建一个时间序列。但是，您的数据集仅包含1天的数据。

构建时间序列后，对其进行预测，提取预测值并将其放入列表中。循环遍历称为“过滤”的所有对象行，定义for循环并将filtered[1, 1]替换为filtered[i, 1]。在循环结束时，将plyr::ldply()列表放入数据框中。

如果您提供更多日期的数据，我可以提供更好的答案。

R中的时间序列预测

1 个答案: