課程大綱

  1. 準備資料
  2. 檢視熱門銷售產品
  3. 關聯規則應用及介紹
  4. 以蔬菜為例應用關聯規則

1. 準備資料 (for Association Rule Analysis) 關聯規則分析

購物籃分析會使用arules這個套件，它要用的資料很容易準備，直接使用交易項目資料，一行程式就可以搞定。
library(arules)  # association rules
## Loading required package: Matrix
## 
## Attaching package: 'arules'
## The following objects are masked from 'package:base':
## 
##     abbreviate, write
library(arulesViz)  # data visualization of association rules
## Loading required package: grid
library(RColorBrewer)  # color palettes for plots
packages = c(
  "dplyr","ggplot2","googleVis","devtools","magrittr","slam","irlba","plotly", "arules","arulesViz","Matrix","recommenderlab")
existing = as.character(installed.packages()[,1])
for(pkg in packages[!(packages %in% existing)]) install.packages(pkg)

1-1. 讀入資料 (read data)

data(Groceries)  # grocery transactions object from arules package
summary(Groceries)
## transactions as itemMatrix in sparse format with
##  9835 rows (elements/itemsets/transactions) and
##  169 columns (items) and a density of 0.02609146 
## 
## most frequent items:
##       whole milk other vegetables       rolls/buns             soda 
##             2513             1903             1809             1715 
##           yogurt          (Other) 
##             1372            34055 
## 
## element (itemset/transaction) length distribution:
## sizes
##    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15 
## 2159 1643 1299 1005  855  645  545  438  350  246  182  117   78   77   55 
##   16   17   18   19   20   21   22   23   24   26   27   28   29   32 
##   46   29   14   14    9   11    4    6    1    1    1    1    3    1 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   1.000   2.000   3.000   4.409   6.000  32.000 
## 
## includes extended item information - examples:
##        labels  level2           level1
## 1 frankfurter sausage meat and sausage
## 2     sausage sausage meat and sausage
## 3  liver loaf sausage meat and sausage
# show the dimensions of the transactions object
print(dim(Groceries))
## [1] 9835  169
print(dim(Groceries)[1])  # 9835 market baskets for shopping trips
## [1] 9835
print(dim(Groceries)[2])  # 169 initial store items  
## [1] 169
inspect(Groceries[1:5])
##     items                     
## [1] {citrus fruit,            
##      semi-finished bread,     
##      margarine,               
##      ready soups}             
## [2] {tropical fruit,          
##      yogurt,                  
##      coffee}                  
## [3] {whole milk}              
## [4] {pip fruit,               
##      yogurt,                  
##      cream cheese ,           
##      meat spreads}            
## [5] {other vegetables,        
##      whole milk,              
##      condensed milk,          
##      long life bakery product}
inspect(Groceries[1])
##     items                
## [1] {citrus fruit,       
##      semi-finished bread,
##      margarine,          
##      ready soups}
# examine frequency for each item with support greater than 0.025
pdf(file="fig_market_basket_initial_item_support.pdf", 
  width = 8.5, height = 11)

itemFrequencyPlot(Groceries, support = 0.025, cex.names=0.8, xlim = c(0,0.3),
  type = "relative", horiz = TRUE, col = "dark red", las = 1,
  xlab = paste("Proportion of Market Baskets Containing Item",
    "\n(Item Relative Frequency or Support)"))

2. 檢視熱門銷售產品 - Top20 熱賣產品

itemFrequencyPlot(Groceries, topN=20, type="absolute",col = "dark red", cex=0.8)

# explore possibilities for combining similar items
df = itemInfo(Groceries)
str(df) # levels 10, 55
## 'data.frame':    169 obs. of  3 variables:
##  $ labels: chr  "frankfurter" "sausage" "liver loaf" "ham" ...
##  $ level2: Factor w/ 55 levels "baby food","bags",..: 44 44 44 44 44 44 44 42 42 41 ...
##  $ level1: Factor w/ 10 levels "canned food",..: 6 6 6 6 6 6 6 6 6 6 ...

3. 關聯規則應用及介紹 - Apriori演算法

關聯規則(A => B)
  • support: A被購買的機率 (A的基礎機率)
  • confidence: A被購買時，B被購買的機率
  • lift: A被購買時，B被購買的機率增加的倍數 (與B的基礎機率相比)
  • 一般來講support、confidence和lift越高的關聯規則越重要
  • support、confidence和lift設的越低(高)，找到的關聯規則越多(少)
  • 門檻值設得越低，找到的規則會越多
  • 建議一開始把標準設低，先找到多一點規則，之後再用subset篩選出特定的規則來看
# aggregate items using the 55 level2 levels for food categories
# to create a more meaningful set of items
groceries <- aggregate(Groceries, itemInfo(Groceries)[["level2"]])  

print(dim(groceries)[1])  # 9835 market baskets for shopping trips
## [1] 9835
print(dim(groceries)[2])  # 55 final store items (categories)  
## [1] 55
各項商品出現在購物欄中的頻率
itemFrequencyPlot(groceries, support = 0.025, cex.names=1.0, xlim = c(0,0.5),
  type = "relative", horiz = TRUE, col = "blue", las = 1,
  xlab = paste("Proportion of Market Baskets Containing Item",
    "\n(Item Relative Frequency or Support)"))

# obtain large set of association rules for items by category and all shoppers
# this is done by setting very low criteria for support and confidence
first.rules <- apriori(groceries, 
  parameter = list(support = 0.001, confidence = 0.05))
## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##        0.05    0.1    1 none FALSE            TRUE       5   0.001      1
##  maxlen target   ext
##      10  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 9 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[55 item(s), 9835 transaction(s)] done [0.00s].
## sorting and recoding items ... [54 item(s)] done [0.00s].
## creating transaction tree ... done [0.00s].
## checking subsets of size 1 2 3 4 5 6 7 8 done [0.02s].
## writing ... [69921 rule(s)] done [0.01s].
## creating S4 object  ... done [0.02s].
print(summary(first.rules))  # yields 69,921 rules... too many
## set of 69921 rules
## 
## rule length distribution (lhs + rhs):sizes
##     1     2     3     4     5     6     7     8 
##    21  1205 10467 23895 22560  9888  1813    72 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   1.000   4.000   4.000   4.502   5.000   8.000 
## 
## summary of quality measures:
##     support           confidence          lift             count        
##  Min.   :0.001017   Min.   :0.0500   Min.   : 0.4475   Min.   :  10.00  
##  1st Qu.:0.001118   1st Qu.:0.2110   1st Qu.: 1.8315   1st Qu.:  11.00  
##  Median :0.001525   Median :0.4231   Median : 2.2573   Median :  15.00  
##  Mean   :0.002488   Mean   :0.4364   Mean   : 2.5382   Mean   :  24.47  
##  3rd Qu.:0.002339   3rd Qu.:0.6269   3rd Qu.: 2.9662   3rd Qu.:  23.00  
##  Max.   :0.443010   Max.   :1.0000   Max.   :16.1760   Max.   :4357.00  
## 
## mining info:
##       data ntransactions support confidence
##  groceries          9835   0.001       0.05
# select association rules using thresholds for support and confidence 
second.rules <- apriori(groceries, 
  parameter = list(support = 0.025, confidence = 0.05))
## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##        0.05    0.1    1 none FALSE            TRUE       5   0.025      1
##  maxlen target   ext
##      10  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 245 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[55 item(s), 9835 transaction(s)] done [0.00s].
## sorting and recoding items ... [32 item(s)] done [0.00s].
## creating transaction tree ... done [0.00s].
## checking subsets of size 1 2 3 4 done [0.00s].
## writing ... [344 rule(s)] done [0.00s].
## creating S4 object  ... done [0.00s].
print(summary(second.rules))  # yields 344 rules
## set of 344 rules
## 
## rule length distribution (lhs + rhs):sizes
##   1   2   3   4 
##  21 162 129  32 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##     1.0     2.0     2.0     2.5     3.0     4.0 
## 
## summary of quality measures:
##     support          confidence           lift            count       
##  Min.   :0.02542   Min.   :0.05043   Min.   :0.6669   Min.   : 250.0  
##  1st Qu.:0.03030   1st Qu.:0.18202   1st Qu.:1.2498   1st Qu.: 298.0  
##  Median :0.03854   Median :0.39522   Median :1.4770   Median : 379.0  
##  Mean   :0.05276   Mean   :0.37658   Mean   :1.4831   Mean   : 518.9  
##  3rd Qu.:0.05236   3rd Qu.:0.51271   3rd Qu.:1.7094   3rd Qu.: 515.0  
##  Max.   :0.44301   Max.   :0.79841   Max.   :2.4073   Max.   :4357.0  
## 
## mining info:
##       data ntransactions support confidence
##  groceries          9835   0.025       0.05
利用散佈圖視覺化關聯規則
  • 有互動圖表的幫助，我們可以把條件放寬，多找一些關聯規則進來
# data visualization of association rules in scatter plot
plot(second.rules,colors=c("red","green"),engine="htmlwidget",
     marker=list(opacity=.6,size=8))
## To reduce overplotting, jitter is added! Use jitter = 0 to prevent jitter.
plot(second.rules,method="matrix",shading="lift",engine="htmlwidget",
     colors=c("red", "green"))

4. 以蔬菜為例應用關聯規則

關聯規則 (A => B)：
  • support: A被購買的機率 (A的基礎機率)
  • confidence: A被購買時，B被購買的機率
  • lift: A被購買時，B被購買的機率增加的倍數 (與B的基礎機率(confidence/lift)相比)
  • count: 這個組合在這個菜籃出現的次數
# select rules with vegetables in consequent (right-hand-side) item subsets
vegie.rules <- subset(second.rules, subset = rhs %pin% "vegetables")
inspect(vegie.rules)  # 41 rules
##      lhs                            rhs             support confidence      lift count
## [1]  {}                          => {vegetables} 0.27300458  0.2730046 1.0000000  2685
## [2]  {poultry}                   => {vegetables} 0.02897814  0.5745968 2.1047148   285
## [3]  {pork}                      => {vegetables} 0.03009659  0.5220459 1.9122238   296
## [4]  {staple foods}              => {vegetables} 0.02613116  0.5160643 1.8903136   257
## [5]  {eggs}                      => {vegetables} 0.03141840  0.4951923 1.8138608   309
## [6]  {games/books/hobby}         => {vegetables} 0.02785968  0.3145809 1.1522918   274
## [7]  {long-life bakery products} => {vegetables} 0.02907982  0.3492063 1.2791227   286
## [8]  {perfumery}                 => {vegetables} 0.03213015  0.4056483 1.4858662   316
## [9]  {beef}                      => {vegetables} 0.04585663  0.5595533 2.0496116   451
## [10] {bags}                      => {vegetables} 0.03141840  0.3175745 1.1632571   309
## [11] {vinegar/oils}              => {vegetables} 0.04199288  0.4666667 1.7093731   413
## [12] {chocolate}                 => {vegetables} 0.03192679  0.2934579 1.0749195   314
## [13] {beer}                      => {vegetables} 0.03406202  0.2189542 0.8020168   335
## [14] {frozen foods}              => {vegetables} 0.04738180  0.4052174 1.4842879   466
## [15] {cheese}                    => {vegetables} 0.05531266  0.4365971 1.5992300   544
## [16] {sausage}                   => {vegetables} 0.07625826  0.4032258 1.4769929   750
## [17] {fruit}                     => {vegetables} 0.10706660  0.4297959 1.5743176  1053
## [18] {non-alc. drinks}           => {vegetables} 0.09456024  0.2974097 1.0893944   930
## [19] {bread and backed goods}    => {vegetables} 0.11621759  0.3363743 1.2321198  1143
## [20] {dairy produce}             => {vegetables} 0.17041179  0.3846683 1.4090180  1676
## [21] {beef,                                                                           
##       dairy produce}             => {vegetables} 0.02989324  0.6074380 2.2250104   294
## [22] {dairy produce,                                                                  
##       vinegar/oils}              => {vegetables} 0.03141840  0.5355286 1.9616103   309
## [23] {dairy produce,                                                                  
##       frozen foods}              => {vegetables} 0.03436706  0.5121212 1.8758704   338
## [24] {cheese,                                                                         
##       fruit}                     => {vegetables} 0.02674123  0.5197628 1.9038613   263
## [25] {bread and backed goods,                                                         
##       cheese}                    => {vegetables} 0.02887646  0.4536741 1.6617821   284
## [26] {cheese,                                                                         
##       dairy produce}             => {vegetables} 0.04219624  0.4987981 1.8270686   415
## [27] {fruit,                                                                          
##       sausage}                   => {vegetables} 0.03426538  0.5290424 1.9378517   337
## [28] {non-alc. drinks,                                                                
##       sausage}                   => {vegetables} 0.03029995  0.4156206 1.5223944   298
## [29] {bread and backed goods,                                                         
##       sausage}                   => {vegetables} 0.04382308  0.4229637 1.5492916   431
## [30] {dairy produce,                                                                  
##       sausage}                   => {vegetables} 0.05266904  0.4905303 1.7967842   518
## [31] {fruit,                                                                          
##       non-alc. drinks}           => {vegetables} 0.04361973  0.4657980 1.7061914   429
## [32] {bread and backed goods,                                                         
##       fruit}                     => {vegetables} 0.05124555  0.4763705 1.7449177   504
## [33] {dairy produce,                                                                  
##       fruit}                     => {vegetables} 0.07869853  0.5032510 1.8433793   774
## [34] {bread and backed goods,                                                         
##       non-alc. drinks}           => {vegetables} 0.04636502  0.3731588 1.3668590   456
## [35] {dairy produce,                                                                  
##       non-alc. drinks}           => {vegetables} 0.06446365  0.4243641 1.5544213   634
## [36] {bread and backed goods,                                                         
##       dairy produce}             => {vegetables} 0.08195221  0.4366197 1.5993128   806
## [37] {dairy produce,                                                                  
##       fruit,                                                                          
##       sausage}                   => {vegetables} 0.02714794  0.5741935 2.1032378   267
## [38] {bread and backed goods,                                                         
##       dairy produce,                                                                  
##       sausage}                   => {vegetables} 0.03284189  0.5135135 1.8809704   323
## [39] {dairy produce,                                                                  
##       fruit,                                                                          
##       non-alc. drinks}           => {vegetables} 0.03304525  0.5183413 1.8986543   325
## [40] {bread and backed goods,                                                         
##       dairy produce,                                                                  
##       fruit}                     => {vegetables} 0.04077275  0.5276316 1.9326840   401
## [41] {bread and backed goods,                                                         
##       dairy produce,                                                                  
##       non-alc. drinks}           => {vegetables} 0.03345196  0.4627286 1.6949480   329
# sort by lift and identify the top 10 rules
top.vegie.rules <- head(sort(vegie.rules, decreasing = TRUE, by = "lift"), 10)
inspect(top.vegie.rules) 
##      lhs                         rhs             support confidence     lift count
## [1]  {beef,                                                                       
##       dairy produce}          => {vegetables} 0.02989324  0.6074380 2.225010   294
## [2]  {poultry}                => {vegetables} 0.02897814  0.5745968 2.104715   285
## [3]  {dairy produce,                                                              
##       fruit,                                                                      
##       sausage}                => {vegetables} 0.02714794  0.5741935 2.103238   267
## [4]  {beef}                   => {vegetables} 0.04585663  0.5595533 2.049612   451
## [5]  {dairy produce,                                                              
##       vinegar/oils}           => {vegetables} 0.03141840  0.5355286 1.961610   309
## [6]  {fruit,                                                                      
##       sausage}                => {vegetables} 0.03426538  0.5290424 1.937852   337
## [7]  {bread and backed goods,                                                     
##       dairy produce,                                                              
##       fruit}                  => {vegetables} 0.04077275  0.5276316 1.932684   401
## [8]  {pork}                   => {vegetables} 0.03009659  0.5220459 1.912224   296
## [9]  {cheese,                                                                     
##       fruit}                  => {vegetables} 0.02674123  0.5197628 1.903861   263
## [10] {dairy produce,                                                              
##       fruit,                                                                      
##       non-alc. drinks}        => {vegetables} 0.03304525  0.5183413 1.898654   325
plot(vegie.rules,method="graph",engine="htmlwidget",itemCol="cyan")
  • 泡泡大小：support: A被購買的機率 (A的基礎機率)
  • 泡泡顏色：lift: A被購買時，B被購買的機率增加的倍數 (與B的基礎機率相比)