Discover Associations Between Products
Sara Marín López, Gabriel Ristow Cidral
24-03-2018
SUMMARY
1.Introduction
Blackwell Electronics’ board of directors is considering acquiring Electronidex, a start-up electronics online retailer. The board of directors has asked us to help them better understand the clientele that Electronidex currently is serving and if it would be an optimal partnership. For this task, we’ll identify purchasing patterns using a market basket analysis.
These tasks were carried out following the next steps:- An exploratory analysis of the transactions product features (With qualitative analysis, cleaning data and creating two different groups: retailers and business)
- Creating rules according to different levels
You can see a section of the used datasets and the complete code we used in R in the Appendix.
2.Results and Recommendations
After a thorough analysis, we have the next conclusions:
- First of all, we would like to clarify that, as the data provided represents just one month of transactions, the results can be biased due to variances between months. Therefore, the conclusion demonstrated in this report should serve just as a reference.
- Analyzing the sales behavior, it gets clear that Electronidex´s clients are both final customers and companies. These groups have completely different shopping behaviors and therefore follow different association patterns. -Electronidex’s final customers tend to prefer products from Apple. The company’s headphones (Active Earpods Apple) appears to be the best product to be recommended. -Electronidex’s corporate clients vary in terms of brand preference. Nevertheless, the monitor ViewSonic seems to be a great product to be recommended.
- Considering that clients don’t buy more than once a month, Electronidex has roughly 9,800 clients, which would enhance Blackwell’s client base.
- The portfolio of both companies is complementary, as Electronidex is more focused on high end products (such as laptops and Desktops) and Blackwell in lower end.
APPENDIX
1. EXPLORATORY ANALYSIS
1.1. Qualitative analysis and cleaning data
We start the exploration thinking about which levels of analysis could be useful for analyzing this dataset. We decided to use two levels:
- Per product: we have 123 different levels
- Per category: we have 18 different levels
-
Identify inconsistencies::
- The “iPhone Charger Cable” would be an Accessory and not a Computer Cord - The “Microsoft Office Home and Student 2016” would be a Software (a new category we created) and not an Accessory - The “Computer Game” would be a Software and not an Accessory - We changed the names of the products: adding the category at the beginning of the word
- We added a level1 (Categories): using the labels
- We removed the empty rows: two of the transactions
1.2. Creating groups: retailers and business
Before starting the creation of rules, we realized that Electronide’s clients are not only final customers, but also companies. In order to split the transactions into the two categories, we follow the next steps.
-
We classify some products as “main products”, other as “complementary products”. We give them scores according to the next categories:
- The transactions with higher scores will be categorized as business, and the transactions with lower scores will be categorized as retailers. For doing this difference more significative, we square the product score as many times as it is repeated in the same transaction.
- This is the final formula for calculating each score -Summation (Product Score ^ repetitions) -Example 1: Acer Desktop + Lenovo Desktop Computer + Intel Desktop + XIBERIA Gaming Headset –> 10^3 + 3 –> 1003 points -Example 2: Acer Aspire + Brother Printer Toner –> 10 + 1 –> 11 points -Example 3: iMac –> 10 points
- The transactions<40 points will be retail and the transactions >40 points will be business.
- Applying this formula, Electrodinex has the following client profile:
2. RETAIL TRANSACTIONS
2.1. Per Categories
Firstly, we visualize the frequency of the different categories. The most purchased categories are Desktops, Laptops and Monitors. 
- -A low level of support (0.025) for creating a lot of rules
- -A medium level of confidence (0.1)
- -Removing the “main products”, Computers and Desktop, from the rhs side of the rules. This is done because main products would influence customers to buy “complementary products”.In a simple example, one person first buys a computer to buy a mouse and not the opposite. Therefore, the products considered “complementary” were included as recommended products after a client purchasing a main product.
This way, we obtain the next 10 top rules (sorted by confidence):
2.2. Per products
Firstly, we visualize the frequency of the different products The most purchased products are 3 Apple products: the active earpods, the laptop MacBookAir and the iMac. It seems that products from Apple appear to be the best seller in the retail channel. 
- -A low level of support (0.025) for creating a lot of rules
- -A medium level of confidence (0.1)
- -Removing the “main products”, Computers and Desktop, from the rhs side of the rules. This is done because main products would influence customers to buy “complementary products”.In a simple example, one person first buys a computer to buy a mouse and not the opposite. Therefore, the products considered “complementary” were included as recommended products after a client purchasing a main product.
This way, we obtain the next 10 top rules (sorted by confidence): 
3. BUSINESS TRANSACTIONS
3.1. Per Categories
Firstly, we visualize the frequency of the different categories. The most purchased categories are again Desktops, Laptops and Monitors. 
- -A low level of support (0.025) for creating a lot of rules
- -A medium level of confidence (0.1)
- -Removing the “main products”, Computers and Desktop, from the rhs side of the rules. This is done because main products would influence customers to buy “complementary products”.In a simple example, one person first buys a computer to buy a mouse and not the opposite. Therefore, the products considered “complementary” were included as recommended products after a client purchasing a main product.
This way, we obtain the next 10 top rules (sorted by confidence):
3.2. Per products
Firstly, we visualize the frequency of the different products The most purchased products are the iMac, the HP Laptop and the Cyberpower Gamer Desktop. 
- -A low level of support (0.025) for creating a lot of rules
- -A medium level of confidence (0.1)
- -Removing the “main products”, Computers and Desktop, from the rhs side of the rules. This is done because main products would influence customers to buy “complementary products”.In a simple example, one person first buys a computer to buy a mouse and not the opposite. Therefore, the products considered “complementary” were included as recommended products after a client purchasing a main product.
This way, we obtain the next 10 top rules (sorted by confidence): 
4. USED DATASETS
A dataset with 9835 transactions and the next products and categories were provided:- Accesories -Microsoft Office Home and Student 2016 -Computer Game -Belkin Mouse Pad -Large Mouse Pad
- Active HeadPhones -Apple Earpods -Monster Beats By Dr Dre -Otium Wireless Sports Bluetooth Headphones -Panasonic In-Ear Headphone -APIE Bluetooth Headphones -Philips Flexible Earhook Headphon
- Computer Cord -HDMI Cable 6ft -Ethernet Cable -Etekcity Power Extension Cord Cable -Audio Cable -VGA Monitor Cable -iPhone Charger Cable -HDMI Adapter -USB Cable -Samsung Charging Cable
- Computer HeadPhones -Zombie Gaming Headset -Logitech ClearChat Headset -Panasonic On-Ear Stereo Headphones RP-HT21 -PC Gaming Headset -Kensington Headphones -Logitech Stereo Headset -Koss Home Headphones -Microsoft Headset -Ailihen Stereo Headphones -XIBERIA Gaming Headset
- Computer Mouse -3 Button -Logitech Wireless -Microsoft Basic Optical -Logitech 3-button -Redragon Gaming -HP Wireless -Generic Black 3-Button -Wireless Portable -Gaming Mouse Professional -Slim Wireless
- Computer Stand -Halter Acrylic Monitor Stand -Height-Adjustable Standing Desk -Multi Media Stan -Halter Mesh Metal Monitor Stand -Full Motion Monitor Mount
- Computer Tablet -iPad -iPad Pro -Fire HD Tablet -Samsung Galaxy Tab -Kindle
- Desktop -Lenovo -iMac -HP -ASUS -Dell -Intel -Acer -Cyberpower Gamer -Dell2
- HardDrive -1TB Portable External -2TB Portable Externa -5TB Desktop -Slim 2TB Portable External -3TB Portable External
- Keyboard -HP USB -Logitech Wireless -Rii LED -Logitech -Backlit LED Gaming -Dell Wired -Apple Wired -Apple Wireless -Apple Magic
- Laptop -LG Touchscreen -Acer Aspire -HP -ASUS Chromebook -Apple Macbook Pro -Apple MacBook Air -Dell -Eluktronics Pro Gaming -Alienware AW17R4-7345SLV-PUS 17" -HP Notebook Touchscreen
- Monitor: -Acer -Acer Aspire -LG -ASUS Chromebook -ASUS 2 -Dell -Samsung -Sceptre -ViewSonic -AOC -HP
- MouseAndKeyboardCombo: -Logitech MK550 Wireless Wave -Logitech Desktop MK120 -Logitech MK270 Wireless -Dell KM117 Wireless -EagleTec Wireless -Microsoft Wireless Comfort -Microsoft Wireless Desktop -Rii LED Gaming -Logitech MK360 Wireless
- Printer -Epson -HP Wireless -Canon Office -Brother -DYMO Label Manker
- PrinterInk -Epson Black -HP Black & Tri-color -Canon -Brother Printer Toner -DYMO Labeling Tape
- SmartHome Device -Apple TV -Google Home -Smart Light Bulb -Fire TV Stick -Roku Express
- Speaker -Cambridge Bluetooth -JBL Splashproof Portable Bluetooth -DOSS Touch Wireless Bluetooth -Logitech Multimedia -Rokono Mini -Cyber Acoustics -Bose Companion -Mackie CR -Sonos
5. COMPLETE CODE
Here you can obtain all the code used for this task.
Includes
#Load Libraries: p_load can install, load, and update packages
if(require("pacman")=="FALSE"){
install.packages("pacman")
}
pacman::p_load(arules, arulesViz, dplyr, ggplot2, treemap,RColorBrewer)
# Load Data
setwd("C:/SARA/Ubiqum/Section2/Task4")
DataTran<- read.transactions("ElectronidexTransactions2017.csv", format = "basket", sep=",", rm.duplicates=TRUE)
# Rename Variables using the same format
namesDataTran<-itemLabels(DataTran)
myLabels<- c("HardDrive_External_1TB_Portable","HardDrive_External_2TBPortable", "ComputerMouse_3Button", "HardDrive_External_3TBPortable","HardDrive_Desktop_5TB","Laptop_Acer_Aspire", "Desktop_Acer", "Monitor_Acer", "ComputerHeadPhones_Ailihen_Stereo", "Laptop_Alienware_AW17R4-7345SLV-PUS17", "Monitor_AOC", "ActiveHeadPhones_APIE_Bluetooh","ActiveHeadPhones_Earpods_Apple", "Laptop_AppleMacBookAir", "Laptop_AppleMacBookPro", "Keyboard_AppleMagic", "SmartHomeDevice_TV_Apple","Keyboard_AppleWired", "Keyboard_AppleWireless", "Monitor_ASUS2", "Laptop_ChromebookASUS", "Desktop_ASUS", "Monitor_ASUS","ComputerCord_AudioCable", "Keyboard_BacklitLED_Gaming", "Accessories_MousePad_Belkin", "Speaker_BoseCompanion", "Printer_Brother", "PrinterInk_Toner_Brother", "Speaker_Bluetooth_Cambridge","PrinterInk_Canon", "Printer_Canon_Office", "Software_ComputerGame", "Speaker_CyberAcoustics", "Desktop_CYBERPOWERGamer", "Desktop_Dell2", "Desktop_Dell", "MouseAndKeyboardCombo_Dell_KM117Wireless", "Laptop_Dell", "Monitor_Dell", "Keyboard_Dell_Wired", "Speaker_DOSS_Touch_Wireless_Bluetooth", "Printer_DYMO_LabelMaker","PrinterInk_DYMOLabelingTape", "MouseAndKeyboardCombo_EagleTec_Wireless", "Laptop_Eluktronics_ProGaming", "PrinterInk_Epson_Black", "Printer_Epson", "ComputerCord_Etekcity_PowerExtension","ComputerCord_EthernetCable", "ComputerTablet_FireHD", "SmartHomeDevice_FireTVStick", "ComputerStand_MonitorFullMotion", "ComputerMouse_GamingProfessional", "ComputerMouse_GenericBlack_3Button", "SmartHomeDevice_GoogleHome", "ComputerStand_HalterAcrylicMonitorStand", "ComputerStand_Halter_MeshMetal_MonitorStand", "ComputerCord_HDMIAdapter", "ComputerCord_HDMICable6ft", "ComputerStand_HeightAdjustable_StandingDesk", "PrinterInk_HP_BlackAndTricolor", "Desktop_HP", "Laptop_HP", "Monitor_HP", "Laptop_HP_Notebook_TouchScreen_PC", "Keyboard_HP_USB", "ComputerMouse_HP_Wireless", "Printer_HP_Wireless", "Desktop_iMac", "Desktop_Intel", "ComputerTablet_iPad", "ComputerTablet_iPadPro", "Accessories_iPhone_Charger_Cable", "Speaker_JBL_Splashproof_Portable_Bluetooth", "ComputerHeadPhones_Kensington", "ComputerTablet_Kindle","ComputerHeadPhones_Koss_Home", "Accessories_LargeMousePad", "Desktop_Lenovo_Computer", "Monitor_LG", "Laptop_LG_Touchscreen", "ComputerMouse_Logitech_3button", "ComputerHeadPhones_HeadSet_Logitech_ClearChat", "MouseAndKeyboardCombo_Logitech_Desktop_MK120", "Keyboard_Logitech", "MouseAndKeyboardCombo_Logitech_MK270_Wireless", "MouseAndKeyboardCombo_Logitech_MK360_Wireless", "MouseAndKeyboardCombo_Logitech_MK550_Wireless_Wave", "Speaker_Logitech_Multimedia", "ComputerHeadPhones_Headset_Logitech_Stereo", "Keyboard_Logitech_Wireless", "ComputerMouse_Logitech_Wireless", "Speaker_Mackie_CR", "ComputerMouse_Microsoft_BasicOptical", "ComputerHeadPhones_HeadSet_Microsoft", "Software_MicrosoftOffice_HomeandStudent2016","MouseAndKeyboardCombo_Microsoft_Wireless_Comfort","MouseAndKeyboardCombo_Microsoft_Wireless_Desktop", "ActiveHeadPhones_MonsterBeats_ByDrDre","ComputerStand_MultiMediaStand","ActiveHeadPhones_Otium_Wireless_Sports_Bluetooth", "ActiveHeadPhones_Panasonic_In-Ear", "ComputerHeadPhones_Panasonic_On-Ear_Stereo_RP-HT21", "ComputerHeadPhones_HeadSet_PC_Gaming", "ActiveHeadPhones_Philips_Flexible_Earhook","ComputerMouse_Redragon_Gaming", "MouseAndKeyboardCombo_Rii_LED_Gaming", "Keyboard_Rii_LED", "Speaker_Rokono_Mini", "SmartHomeDevice_RokuExpress", "ComputerCord_Samsung_ChargingCable", "ComputerTablet_SamsungGalaxy", "Monitor_Samsung", "Monitor_Sceptre","HardDrive_External_Slim_2TB_Portable", "ComputerMouse_Slim_Wireless", "SmartHomeDevice_SmartLightBulb", "Speaker_Sonos", "ComputerCord_USB_Cable", "ComputerCord_VGA_MonitorCable","Monitor_ViewSonic", "ComputerMouse_Wireless_Portable", "ComputerHeadPhones_HeadSet_XIBERIA_Gaming", "ComputerHeadPhones_HeadSet_Zombie_Gaming")
itemLabels(DataTran)<-myLabels
# Create Level
myLevel1<-as.factor(sub("\\_.*", "", myLabels))
myLevel1
# Add Level
DataTran@itemInfo$level1<-myLevel1
# Remove empty rows
DataTran<-DataTran[which(size(DataTran)!= 0)]Creating groups (Retail / Business)
# We have 9833 transactions
dim(DataTran)[1]
# Create Scores
scores<-c()
for (i in 1:9833){
PAccessories=0
PActiveHeadPhones=0
PComputerCord=0
PComputerHeadPhones=0
PComputerMouse=0
PComputerStand=0
PComputerTablet=0
PDesktop_Laptop=0
PHardDrive=0
PKeyboard=0
PMonitor=0
PMouseAndKeyboardCombo=0
PPrinter=0
PPrinterInk=0
PSmartHomeDevice=0
PSoftware=0
PSpeaker=0
print("New Transaction")
print(i)
if (length(grep("Accessories_", LIST(DataTran)[[i]]))>=1){
PAccessories<-(3^length(grep("Accessories_", LIST(DataTran)[[i]])))
# print("PAccessories")
# print(PAccessories)
}
if (length(grep("ActiveHeadPhones_", LIST(DataTran)[[i]]))>=1){
PActiveHeadPhones<-(3^length(grep("ActiveHeadPhones_", LIST(DataTran)[[i]])))
# print("PActiveHeadPhones")
# print(PActiveHeadPhones)
}
if (length(grep("ComputerCord_", LIST(DataTran)[[i]]))>=1){
PComputerCord<-(3^length(grep("ComputerCord_", LIST(DataTran)[[i]])))
# print("PComputerCord")
# print(PComputerCord)
}
if (length(grep("ComputerHeadPhones_", LIST(DataTran)[[i]]))>=1){
PComputerHeadPhones<-(3^length(grep("ComputerHeadPhones_", LIST(DataTran)[[i]])))
# print("PComputerHeadPhones")
# print(PComputerHeadPhones)
}
if (length(grep("ComputerMouse_", LIST(DataTran)[[i]]))>=1){
PComputerMouse<-(3^length(grep("ComputerMouse_", LIST(DataTran)[[i]])))
# print("PComputerMouse")
# print(PComputerMouse)
}
if (length(grep("ComputerStand_", LIST(DataTran)[[i]]))>=1){
PComputerStand<-(3^length(grep("ComputerStand_", LIST(DataTran)[[i]])))
# print("PComputerStand")
# print(PComputerStand)
}
if (length(grep("ComputerTablet_", LIST(DataTran)[[i]]))>=1){
PComputerTablet<-(3^length(grep("ComputerTablet_", LIST(DataTran)[[i]])))
# print("PComputerTablet")
# print(PComputerTablet)
}
if (length(grep("Desktop_", LIST(DataTran)[[i]])>=1) | length(grep("Laptop_", LIST(DataTran)[[i]])>=1)) {
PDesktop_Laptop<-(10^(length(grep("Desktop_", LIST(DataTran)[[i]]))
+ length(grep("Laptop", LIST(DataTran)[[i]]))))
# print("PDesktop")
# print(PDesktop)
}
if (length(grep("HardDrive_", LIST(DataTran)[[i]]))>=1){
PHardDrive<-(3^length(grep("HardDrive_", LIST(DataTran)[[i]])))
# print("PHardDrive")
# print(PHardDrive)
}
if (length(grep("Keyboard_", LIST(DataTran)[[i]]))>=1){
PKeyboard<-(3^length(grep("Keyboard_", LIST(DataTran)[[i]])))
# print("PKeyboard")
# print(PKeyboard)
}
if (length(grep("Monitor_", LIST(DataTran)[[i]]))>=1){
PMonitor<-(3^length(grep("Monitor_", LIST(DataTran)[[i]])))
# print("PMonitor")
# print(PMonitor)
}
if (length(grep("MouseAndKeyboardCombo_", LIST(DataTran)[[i]]))>=1){
PMouseAndKeyboardCombo<-(3^length(grep("MouseAndKeyboardCombo_", LIST(DataTran)[[i]])))
# print("PMouseAndKeyboardCombo")
# print(PMouseAndKeyboardCombo)
}
if (length(grep("Printer_", LIST(DataTran)[[i]]))>=1){
PPrinter<-(3^length(grep("Printer_", LIST(DataTran)[[i]])))
# print("PPrinter")
# print(PPrinter)
}
if (length(grep("PrinterInk_", LIST(DataTran)[[i]]))>=1){
PPrinterInk<-(1^length(grep("PrinterInk_", LIST(DataTran)[[i]])))
# print("PPrinterInk")
# print(PPrinterInk)
}
if (length(grep("SmartHomeDevice_", LIST(DataTran)[[i]]))>=1){
PSmartHomeDevice<-(3^length(grep("SmartHomeDevice_", LIST(DataTran)[[i]])))
# print("PSmartHomeDevice")
# print(PSmartHomeDevice)
}
if (length(grep("Software_", LIST(DataTran)[[i]]))>=1){
PSoftware<-(3^length(grep("Software_", LIST(DataTran)[[i]])))
# print("PSoftware")
# print(PSoftware)
}
if (length(grep("Speaker_", LIST(DataTran)[[i]]))>=1){
PSpeaker<-(3^length(grep("Speaker_", LIST(DataTran)[[i]])))
# print("PSpeaker")
# print(PSpeaker)
}
finalpuntuation<-sum(PAccessories, PActiveHeadPhones, PComputerCord, PComputerHeadPhones,
PComputerMouse, PComputerStand, PComputerTablet, PDesktop_Laptop, PHardDrive,
PKeyboard, PMonitor, PMouseAndKeyboardCombo, PPrinter,
PPrinterInk, PSmartHomeDevice,PSoftware, PSpeaker)
# print(finalpuntuation)
#scores<-union(scores,finalpuntuation)
scores<-rbind(scores, finalpuntuation)
# print("scores")
# print(scores)
}
# Create groups
numbers_small<-c()
for(i in 1:9833)
{ if(scores[i, ]<40){
numbers_small<-rbind(numbers_small, i)
}
}
DataTran_Small<-DataTran[numbers_small]
numbers_high<-c()
for(i in 1:9833)
{ if(scores[i, ]>=40){
numbers_high<-rbind(numbers_high, i)
}
}
DataTran_High<-DataTran[numbers_high]Retail Transactions: analysis per category
# We have 18 different product-types
print(levels(itemInfo(DataTran)[["level1"]]))
DataTran_ProdTypes<- aggregate(DataTran, itemInfo(DataTran)[["level1"]])
DataTran_ProdTypes_Small<-DataTran_ProdTypes[numbers_small]
DataTran_ProdTypes_High<-DataTran_ProdTypes[numbers_high]
#### Frequency Plot ####
itemFrequencyPlot(DataTran_ProdTypes_Small, topN = 18, type ="relative", horiz=TRUE, cex.names=0.8,
xlim=c(0,0.4), las=1, col="dark blue",
xlab = "% Transactions containing Products")
#### TreeMap ####
Labels_Small_ProdTypes<-levels(itemInfo(DataTran)[["level1"]])
Size_Small_ProdTypes<-c()
for (i in 1: length(levels(itemInfo(DataTran)[["level1"]]))) {
Size_Small_ProdTypes<-rbind(Size_Small_ProdTypes, itemFrequency(DataTran_ProdTypes_Small)[i] * 5308)
}
Size_Small_ProdTypes<-as.vector(Size_Small_ProdTypes)
dfsmallprod<-data.frame(Labels_Small_ProdTypes, Size_Small_ProdTypes)
treemap(dfsmallprod,index="Labels_Small_ProdTypes", vSize="Size_Small_ProdTypes",title="",palette="Dark2",border.col="#FFFFFF")
#### Rules ####
"%ni%" <- Negate("%in%")
# Rules with No Laptops/ Desktops on the right side
Rules_Small_ProductType<-apriori(DataTran_ProdTypes_Small,
parameter = list(support=0.025, confidence=0.1, minlen=2),
appearance = list(rhs=c("Accessories", "ActiveHeadPhones",
"ComputerCord","ComputerHeadPhones","ComputerMouse",
"ComputerStand", "ComputerTablet","HardDrive", "Keyboard",
"Monitor","MouseAndKeyboardCombo","Printer", "PrinterInk",
"SmartHomeDevice", "Software", "Speaker")))
norhs_laptpos_desktop<-myLabels[!myLabels %in%
# Inspect
inspect(sort(Rules_Small_ProductType, by='confidence', decreasing = T)[1:10])
# Rules Plots
plot(sort(Rules_Small_ProductType, by='confidence', decreasing = T)[1:10], method="grouped")
plot(sort(Rules_Small_ProductType, by='confidence', decreasing = T)[1:10], method="graph", engine="interactive")Retail Transactions: analysis per product
#### Frequency Plot ####
itemFrequencyPlot(DataTran_Small, topN = 18, type ="relative", horiz=TRUE, cex.names=0.8,
xlim=c(0,0.2), las=1, col="dark blue",
xlab = "% Transactions containing Items")
#### TreeMap ####
Labels_Small_Items<-itemInfo(DataTran_Small)$level1
Labels_Small_ItemsProd<-itemInfo(DataTran_Small)$labels
Size_Small_Items<-c()
for (i in 1: length(itemInfo(DataTran_Small)[["labels"]])) {
Size_Small_Items<-rbind(Size_Small_Items, itemFrequency(DataTran_Small)[i] * 5308)
}
Size_Small_Items<-as.vector(Size_Small_Items)
Labels_Small_ItemsProd<-c(itemInfo(DataTran_Small)$level1)
dfsmallitem<-data.frame(itemInfo(DataTran_Small)$level1, itemInfo(DataTran_Small)$labels, Size_Small_Items)
treemap(dfsmallitem,index=c("itemInfo.DataTran_Small..level1","itemInfo.DataTran_Small..labels"), vSize="Size_Small_Items",title="",palette="Dark2",border.col="#FFFFFF")
#### Rules ####
# No Laptops/ Desktops on the right side
norhs_laptpos_desktop<-myLabels[!myLabels %in%
c("Laptop_Acer_Aspire", "Laptop_Alienware_AW17R4-7345SLV-PUS17",
"Laptop_AppleMacBookAir","Laptop_AppleMacBookPro",
"Laptop_ChromebookASUS", "Laptop_Dell",
"Laptop_Eluktronics_ProGaming", "Laptop_HP",
"Laptop_HP_Notebook_TouchScreen_PC", "Laptop_LG_Touchscreen",
"Desktop_Acer" ,"Desktop_ASUS", "Desktop_CYBERPOWERGamer",
"Desktop_Dell","Desktop_Dell2", "Desktop_HP", "Desktop_iMac",
"Desktop_Intel","Desktop_Lenovo_Computer")]
RulesSmallItems<-apriori(DataTran_Small, parameter = list(support=0.001, confidence=0.1, minlen=2),
appearance = list(rhs=norhs_laptpos_desktop))
inspect(sort(RulesSmallItems, by='confidence', decreasing = T)[1:10])
# Rules Plots
plot(sort(RulesSmallItems, by='confidence', decreasing = T)[1:10], method="grouped")
plot(sort(RulesSmallItems, by='confidence', decreasing = T)[1:10], method="graph", engine="interactive")Business Transactions: analysis per category
#### Frequency Plot ####
itemFrequencyPlot(DataTran_ProdTypes_High, topN = 18, type ="relative", horiz=TRUE, cex.names=0.8,
xlim=c(0,1), las=1, col="dark blue",
xlab = "% Transactions containing Products")
#### TreeMap ####
Labels_High_ProdTypes<-levels(itemInfo(DataTran)[["level1"]])
Size_High_ProdTypes<-c()
for (i in 1: length(levels(itemInfo(DataTran)[["level1"]]))) {
Size_High_ProdTypes<-rbind(Size_High_ProdTypes, itemFrequency(DataTran_ProdTypes_High)[i] * 5308)
}
Size_High_ProdTypes<-as.vector(Size_High_ProdTypes)
dfhighprod<-data.frame(Labels_High_ProdTypes, Size_High_ProdTypes)
treemap(dfhighprod,index="Labels_High_ProdTypes", vSize="Size_High_ProdTypes",title="",palette="Dark2",border.col="#FFFFFF")
#### Rules ####
"%ni%" <- Negate("%in%")
# Rules with No Laptops/ Desktops on the right side
Rules_High_ProductType<-apriori(DataTran_ProdTypes_High,
parameter = list(support=0.025, confidence=0.1, minlen=2),
appearance = list(rhs=c("Accessories", "ActiveHeadPhones",
"ComputerCord","ComputerHeadPhones","ComputerMouse",
"ComputerStand", "ComputerTablet","HardDrive",
"Keyboard", "Monitor","MouseAndKeyboardCombo","Printer",
"PrinterInk", "SmartHomeDevice", "Software", "Speaker")))
inspect(sort(Rules_High_ProductType, by='confidence', decreasing = T)[1:10])
# Rules Plots
plot(sort(Rules_High_ProductType, by='confidence', decreasing = T)[1:10], method="grouped")
plot(sort(Rules_High_ProductType, by='confidence', decreasing = T)[1:10], method="graph", engine="interactive")Business Transactions: analysis per product
#### Frequency Plot ####
itemFrequencyPlot(DataTran_High, topN = 18, type ="relative", horiz=TRUE, cex.names=0.8,
xlim=c(0,0.5), las=1, col="dark blue",
xlab = "% Transactions containing Items")
#### Rules ####
# No Laptops/ Desktops on the right side
norhs_laptpos_desktop2<-myLabels[!myLabels %in%
c("Laptop_Acer_Aspire",
"Laptop_Alienware_AW17R4-7345SLV-PUS17",
"Laptop_AppleMacBookAir","Laptop_AppleMacBookPro",
"Laptop_ChromebookASUS", "Laptop_Dell",
"Laptop_Eluktronics_ProGaming","Laptop_HP",
"Laptop_HP_Notebook_TouchScreen_PC",
"Laptop_LG_Touchscreen","Desktop_Acer",
"Desktop_ASUS", "Desktop_CYBERPOWERGamer",
"Desktop_Dell","Desktop_Dell2", "Desktop_HP",
"Desktop_iMac", "Desktop_Intel",
"Desktop_Lenovo_Computer")]
RulesHighItems<-apriori(DataTran_High, parameter = list(support=0.001, confidence=0.1, minlen=2),
appearance = list(rhs=norhs_laptpos_desktop))
inspect(sort(RulesHighItems, by='confidence', decreasing = T)[1:10])
# Rules Plots
plot(sort(RulesHighItems, by='confidence', decreasing = T)[1:10], method="grouped")
plot(sort(RulesHighItems, by='confidence', decreasing = T)[1:10], method="graph", engine="interactive")