Application of an XMLbased Taxonomy to Retail Industry

1
Application of an XML-based Taxonomy to Retail
Industry

• Intelligent Recommender for Supermarket Shopping

2
Outline

• Introduction (by Vincent)
• Taxonomy (by Sheng)
• Promotion Rules (by Judy)
• Recommendation Rules (by Judy)
• Membership Rules (by Xianrong)
• Discount Rules (by Xianrong)
• Customer Shopping List (by Vincent)
• Total Amount for Current Purchase (by Vincent)

3
Goals of the project

• Supermarket Business Recommend System
• Functionalities of Our System
• Manage memberships
• Manage promotions, special discounts
• Give the discount amount for a product
• Give a recommendation list depending on the
  customer
• Give the total amount of the purchased product

4
Ontology Domain Analysis

• Customers
• Products

5
Ontology Domain Analysis

• Customer

career gender age boughtProducts income hobbies ag
eofChildren totalAmount
6
Ontology Domain Analysis

• Product

inSeason
foodFeature
price expiryDate degreeofPerishablity
7
Ontology Domain Analysis

• Product
• a. Clothes

8
Ontology Domain Analysis

• Product
• b. Food

9
Ontology Domain Analysis

• Product
• c. Entertainment

10
Ontology Domain Analysis

• Product
• d. Sports

11
Ontology Domain Analysis

• Product
• e. Cosmetics

12
Ontology Creating Classes

• All class hierarchy

13
Ontology Creating Properties

• Properties

14
Ontology Creating Individuals

• Customer

15
Ontology Creating Individuals

• Product

16
Ontology Source Code

• Code---gtshow RDF/XML source code

17
Ontology Convert to RDF

• Transformation process

OWL Source code
RDF Files
Class relationships
J2SE JDOM
18
Ontology Convert to RDF

• OWLTORDF interface

19
Knowledge Base Overview
Time
Special Events
Product Attributes
Product Promotion
Customer Attributes
Product Recommendation
Customer Membership
Final Discount
Customer Makes Decisions
Customer Preference
Customer Shopping List
Trolley Computation
20
Promotion Rules
21
Promotion Rules (cont.)

• rule (1)
• If there is a special event, all the products
  will be set to a 50 discount.
• promotion(id-gt?Product_IdProduct
  discountAmount-gt 0.5Real)- special_event(special
  _event_discount-gt0.5Real),
• product(id-gt?Product_Id!?).

22
Promotion Rules (cont.)

• rule (2)
• If a kind of food is imperishable, and it will
  expire in one week, this product will be reduced
  to a 10 discount.
• promotion(id-gt?Product_IdProduct
  discountAmount-gt0.1Real) - product(id-gt?Product
  _IdFood expiryDate-gtoneweek degreeofPerishablit
  y-gtnone!?),
• special_event(special_event_discount-gt0Real).
• rule (3) (4)

23
Promotion Rules (cont.)

• rule (5)
• If a kind of Cosmetics is imperishable, and it
  will expire in one month, this product will be
  reduced to a 50 discount.
• promotion(id-gt?Product_IdProduct
  discountAmount-gt0.5Real) -
• product(id-gt?Product_IdCosmetics
  expiryDate-gtonemonth degreeofPerishablity-gtnone!?
  ),
• special_event(special_event_discount-gt0Real).
• rule (6)(7)(8)

24
Promotion Rules (cont.)

• rule (9)
• When it is out of season, set womens clothes to
  a 50 discount, mens to a 30 discount and kids
  to a 40 discount.
• promotion(id-gt?Product_IdProduct
  discountAmount-gt0.5Real) -
• product(id-gt?Product_Id WomenClothes
  inSeason-gtfalse!?) ,
• special_event(special_event_discount-gt0Real).
• rule (10)(11)

25
Promotion Rules (Demo)
26
Recommendation Rules
27
Recommendation Rules (cont.)

• member rule identify whether an element is in
  a list
• member(?Product_IdProduct, ?Product_IdProduct
  ?CustomerBoughtProducts_Tail).
• member(?Product_IdProduct,?CustomerBoughtProduct
  s_HeadProduct ?CustomerBoughtProducts_Tail)- n
  otEqual(?Product_IdProduct,?CustomerBoughtProduct
  s_HeadProduct),
• member(?Product_IdProduct,? CustomerBoughtProdu
  cts_Tail).

28
Recommendation Rules (cont.)

• check rule check an element in a list is less
  than 3 or not
• check (?Customer_Ageofchildren_Head
  ?Customer_Ageofchildren_Tail)- lessThanOrEqual(
  ?Customer_Ageofchildren_HeadInteger, 3Integer).
• check (?Customer_Ageofchildren_Head
  ?Customer_Ageofchildren_Tail )-
  greaterThan(?Customer_Ageofchildren_HeadInteger,
  3Integer),
• check(?Customer_Ageofchildren_Tail).

29
Recommendation Rules (cont.)

• rule (14)
• If a product is on sale and a customer bought
  this product in his previous purchases, recommend
  this product to this customer.
• recommendList(customerid-gt?Customer_Id
  customername-gt?Customer_Name productid-gt?Product_
  IdProduct productname-gt?Product_Name
  discountamount-gt?Discount_AmountReal
  originalprice-gt?Original_PriceReal)-
• product(id-gt?Product_IdProduct
  name-gt?Product_Name price-gt?Original_PriceReal!?
  ),
• promotion(id-gt?Product_IdProduct
  discountAmount-gt?Discount_AmountReal),
• customer(id-gt?Customer_Id name-gt?Customer_Name
  boughtProducts-gt?CustomerBoughtProducts!?),
• member(?Product_IdProduct,?CustomerBoughtProduct
  s).

30
Recommendation Rules (cont.)

• rule (15)
• This is a rule based on a customers hobbies. If
  a customer is fond of swimming and he has never
  bought a kind of swimwear clothes in his previous
  purchases, recommend promoted clothes suitable
  for doing such an exercise to him.
• recommendList(customerid-gt?Customer_Id
  customername-gt?Customer_Name productid-gt?Product_
  IdProduct productname-gt?Product_Name
  discountamount-gt?Discount_AmountReal
  originalprice-gt?Original_PriceReal)-
• product(id-gt?Product_IdSwimwear
  name-gt?Product_Name price-gt?Original_PriceReal!?
  ),
• promotion(id-gt?Product_IdProduct
  discountAmount-gt?Discount_AmountReal),
• customer(id-gt?Customer_Id name-gt?Customer_Name
  hobbies-gtswimming boughtProducts-gt?CustomerBought
  Products!?),
• naf(member(?Product_IdProduct,?CustomerBoughtPro
  ducts)).
• rule (16)Carrer rule(17)AgeofChildren

31
Recommendation Rules (cont.)

• rule (19)
• If a customer is an office white-collar and a
  woman, recommend lady suits and lady dresses to
  her, no matter whether she bought such clothes in
  her previous purchases or whether such clothes
  are on sale.
• recommendList(customerid-gt?Customer_Id
  customername-gt?Customer_Name productid-gt?Product_
  IdProduct productname-gt?Product_Name
  discountamount-gt0Real originalprice-gt?Original_P
  riceReal)-
• product(id-gt?Product_IdWomenDress
  name-gt?Product_Name price-gt?Original_PriceReal!?
  ),
• naf(promotion(id-gt?Product_IdProduct!?)),
• customer(id-gt?Customer_Id name-gt?Customer_Name
  career-gtofficeWhiteCollar gender-gtfemale!?).
• rule (18)hobbies rule(20)AgeofChildren

32
Recommendation Rules Demo
33
Membership Rules
Three Star Member
Total Amount 2000 CAD
Two Star Member
Total Amount 1000 CAD
One Star Member
34
Membership Rules Rule (12)

• membership(id-gt?Customer_Id membershiplevel-gttwoS
  tarMember)-
• customer(id-gt?Customer_Id membership-gtoneStarMemb
  er totalAmount-gt?Customer_TotalAmountReal!?),
• greaterThanOrEqual(?Customer_TotalAmount,
  1000Real).

35
Membership Rules Rule (13)

• membership(id-gt?Customer_Id membershiplevel-gtthre
  eStarMember)-
• customer(id-gt?Customer_Id membership-gttwoStarMemb
  er totalAmount-gt?Customer_TotalAmountReal!?),
• greaterThanOrEqual(?Customer_TotalAmount,
  2000Real).

36
Final Discount Rules Structural Overview
Final discount
Product based discount
Membership based discount
37

Final Discount Rules Formula

• 1-(1-p)(1-m)
• 1-(1-p-mpm)
• 1-1pm-pm
• pm-pm

Product based Discount
Membership Based Discount
38
Final Discount Rules Three Steps

• pm? m-pm? pm-pm
• multiply (?YReal, ?pReal, mReal),
• subtract (?XReal, mReal, ?YReal),
• add(?Final_DiscountReal, ?pReal, ?XReal).

39
Final Discount Rules Rule (21)

• discount(customerid-gt?Customer_Id
  customername-gt?Customer_Name productid-gt?Product_
  IdProduct
• productname-gt?Product_Name originalprice-gt?Origin
  al_PriceReal
• discountamount-gt?Final_Discount_AmountReal)-
• customer(id-gt?Member_Id membership-gtoneStarMember
  !?),
• recommendList(customerid-gt?Customer_Id
  customername-gt?Customer_Name productid-gt?Product_
  IdProduct productname-gt?Product_Name
  originalprice-gt?Original_PriceReal
  discountamount-gt?Discount_AmountReal),
• equal(?Member_Id, ?Customer_Id ),
• Membership
  based discount for one stars member is 0.1
• multiply (?YReal, ?Discount_AmountReal,
  0.1Real),
• subtract (?XReal, 0.1Real, ?YReal),
• add(? Final_Discount_AmountReal,
  ?Discount_AmountReal, ?XReal).

40
Final Discount Rules Rule (22)

• discount(customerid-gt?Customer_Id
  customername-gt?Customer_Name productid-gt?Product_
  Id Product
• productname-gt?Product_Name originalprice-gt?Origin
  al_PriceReal
• discountamount-gt?Final_Discount_AmountReal)-
• customer(id-gt?Member_Id membership-gttwoStarMember
  !?),
• recommendList(customerid-gt?Customer_Id
  customername-gt?Customer_Name productid-gt?Product_
  IdProduct
• productname-gt?Product_Name originalprice-gt?Origin
  al_PriceReal discountamount-gt?Discount_AmountRe
  al),
• Membership
  based discount for two stars member is 0.2
• equal(?Member_Id, ?Customer_Id ),
• multiply (?YReal, ?Discount_AmountReal,
  0.2Real),
• subtract (?XReal, 0.2Real, ?YReal),
• add(?Final_Discount_AmountReal,
  ?Discount_AmountReal, ?XReal).

41
Final Discount Rules Rule (23)

• discount(customerid-gt?Customer_Id
  customername-gt?Customer_Name productid-gt?Product_
  Id Product
• productname-gt?Product_Name originalprice-gt?Origin
  al_PriceReal
• discountamount-gt? Final_Discount_AmountReal)-
• customer(id-gt?Member_Id membership-gtthreeStarMemb
  er!?),
• recommendList(customerid-gt?Customer_Id
  customername-gt?Customer_Name productid-gt?Product_
  IdProduct
• productname-gt?Product_Name originalprice-gt?Origin
  al_PriceReal discountamount-gt?Discount_AmountRe
  al),
• equal(?Member_Id, ?Customer_Id ),
• Membership
  based discount for three stars member is 0.4
• multiply (?YReal, ?Discount_AmountReal,
  0.4Real),
• subtract (?XReal, 0.4Real, ?YReal),
• add(?Final_Discount_AmountReal,
  ?Discount_AmountReal, ?XReal).

42
Final Discount Rules Demo