Part 15: Knowledge-Based Recommender Systems. Francesco Ricci

Transcription

1 Part 15: Knowledge-Based Recommender Systems Francesco Ricci

2 Content p Knowledge-based recommenders: definition and examples p Case-Based Reasoning p Instance-Based Learning p A recommender system exploiting a simple case model (the product is a case) p A more complex CBR recommender system for travel planning 2

3 Core Recommendation Techniques U is a set of users I is a set of items/products [Burke, 2007] 3

4 Knowledge Based Recommender p Suggests products based on inferences about a user s needs and preferences p Functional knowledge: about how a particular item meets a particular user need p The user model can be any knowledge structure that supports this inference n A query, i.e., the set of preferred features for a product n A case (in a case-based reasoning system) n An adapted similarity metric (for matching) n A part of an ontology p There is a large use of domain knowledge encoded in a knowledge representation language/approach. 4

5 ActiveBuyersGuide 5

6 Wizard: My Product Advisor Possible user's requests The system decides what the wizard says

7 7

8 8

9 9

10 10

11 Trip.com 11

12 12

13 13

14 14

15 Matching in TripleHop Example: TripleHop Catalogue of Destinations C-UM:00341 activities relaxing lying on a beach shopping sitting in cafes constraint meat = beef budget = 200 matching [Delgado and Davidson, 2002] 15

16 TripleHop and Content-Based RS p p p p p The content (destination description) is exploited in the recommendation process A classical Content-Based method would have used a simpler content model,e.g., keywords or TF-IDF Here a more complex knowledge structure a tree of concepts is used to model the product (and the query) The query is the user model and it is acquired every time the user asks for a new recommendation - (not exactly, more details later) n Stress on ephemeral needs rather than building a persistent user model Typical in Knowledge-Based RS, they are more focused on ephemeral users because Collaborative Filtering and Content-Based methods cannot cope with that users. 16

17 Learning User Profile: query mining C-UM:00341 Crete C-UM:00357 activitie s constraint activitie s constraint relaxing shopping budget = 200 relaxing adventure lying on a beach sitting in cafes Old query user model activitie s meat = beef C-UM:00357bis constraint lying on a beach hiking meat = pork new user request lying on a beach relaxing sitting in cafes shopping hiking adventure meat = pork budget = 200 new user request, as computed by the systems. Shadowed means less important. 17

18 Query Augmentation p Personalization in search is not only information filtering p Query augmentation: when a query is entered it can be compared against contextual and individual information to refine the query n Ex1: If the user is searching for a restaurant and enter a keyword Thai then the query can be augmented to Thai food (See Part 8 Query expansion based on co-occurrence analysis in the corpus of documents) n Ex2: If the query Thai food does not retrieve any restaurant the query can be refined to Asian food n Ex3: If the query Asian food retrieves too many restaurant, and the user searched in the past for Chinese food the query can be refined to Chinese food. 18

19 Query Augmentation in TripleHop 1. The current query is compared with previous queries of the same user 2. Preferences expressed in past (similar) queries are identified 3. A new query is built by combining the short term preferences contained in the query with the inferred preferences extracted from the persistent user model (past queries) 4. When the query is matched against an item (destination) if two destinations have the same degree of matching for the explicit preferences then the inferred preferences are used to break the tie p This is another example of the cascade approach n the two combined RS are based on the same knowledge but with two definitions of the user model. 19

20 What is Case Based Reasoning? p A case-based reasoner solves new problems by adapting solutions that were used to solve old problems (Riesbeck & Shank 1989) p CBR problem solving process: n store previous experiences (cases) in memory n to solve new problems p Retrieve form the memory similar experience about similar situations p Reuse the experience in the context of the new situation: complete or partial reuse, or adapt according to differences p Store new experience in memory (learning) [Aamodt and Plaza, 1994] 20

21 Case-Based Reasoning [Aha, 1998] 21

22 CBR Assumption p New problem can be solved by n retrieving similar problems n adapting retrieved solutions p Similar problems have similar solutions P P P? P P P P P P S S S X S S S S S S 22

23 Examples of CBR p Classification: The patient s ear problems are like this prototypical case of otitis media p Compiling solutions: Patient N s heart symptoms can be explained in the same way as previous patient D s p Assessing values: My house is like the one that sold down the street for $250,000 but has a better view p Justifying with precedents: This Missouri case should be decided just like Roe v. Wade where the court held that a state s limitations on abortion are illegal p Evaluating options: If we attack Cuban/Russian missile installations, it would be just like Pearl Harbor 23

24 Instance-based learning Lazy Learning p One way of solving tasks of approximating discrete or real valued target functions p Have training examples: (x n, f(x n )), n=1,...,n p Key idea: n just store the training examples n when a test example is given then find the closest matches n use the closest matches to guess the value of the target function on the test example. 24

25 The distance between examples p We need a measure of distance (or similarity) in order to know who are the neighbours p Assume that we have T attributes for the learning problem. Then one example point x has elements x t R, t=1,,t p The distance between two points x and y is often defined as the Euclidean distance: d( x, y) = T t= 1 [ x t y t 2 ] 25

26 no one no two yes three yes four no five six yes no seven Eight?? 26

27 Training data Number Lines Line types Rectangles Colours Mondrian? No No Yes Yes No Yes No Test instance Number Lines Line types Rectangles Colours Mondrian?

28 Lines LinesT Rect Colors Class Distance to test Train no 3,32 Train yes 2,83 Train yes 2,45 Train no 2,65 Train yes 2,65 Train no 5,20 Feature values are not normalized test Train1-0,32 0,32-0,11 0,06 no 0,80 Train2-0,08 0,32-0,21-0,28 yes 0,52 Train3-0,08-0,16-0,11-0,28 yes 0,69 Train4-0,08-0,16 0,08 0,06 no 0,77 Train5 0,16-0,16-0,11 0,39 yes 0,86 Train6 0,40-0,16 0,47 0,06 no 0,76 test 0,40 0,32-0,02-0,28 x' = (x avg(x))/4*stdev(x)), where x is a feature value of the feature X Feature values are normalized What is the difference between this feature values normalization and vector normalization in IR? 28

29 Example of CBR Recommender System p Entree is a restaurant recommender system it finds restaurants: 1. matching some user goals (case features) 2. or similar to restaurants the user knows and likes 29

30 The Product is the Case p In Entrée a case is a restaurant the case is the product p The problem component is the description of the restaurant given by the user p The user will input a partial description of it this is the only difficulty p The solution part of the case is the restaurant itself i.e. the name of the restaurant p The assumption is that the needs of the user can be modeled as the features of the product description. 30

31 Partial Match p In general, only a subset of the preferences will be matched in the recommended restaurant. 31

32 Nearest Neighbor 32

33 Recommendation in Entree p The system first selects from the database the set of all restaurants that satisfy the largest number of logical constraints generated by considering the input features type and value p If necessary, implicitly relaxes the lowest important constraints until some restaurants could be retrieved n Typically the relaxation of constraints will produce many restaurants in the result set p Sorts the retrieved cases using a similarity metric this takes into account all the input features. 33

34 Similarity in Entree p This similarity metric assumes that the user goals, corresponding to the input features (or the features of the source case), could be sorted to reflect the importance of such goals from the user point of view p Hence the global similarity metric (algorithm) sorts the products first with respect the most important goal and then iteratively with respect to the remaining goals (multi-level sort) p Attention: it does not works as a maximization of a Utility-Similarity defined as the sum of local utilities. 34

35 Example Restaurant Price Cusine Atmosphere Dolce 10 A A Gabbana 12 B B p p p p If the user query q is: price=9 AND cusine=b AND Atm=B And the weights (importance) of the features is: 0.5 price, 0.3 Cusine, and 0.2 Atmosphere The Entrée will suggest Dolce first (and then Gabbana) A more traditional CBR system will suggest Gabbana because the similarities are (30 is the price range): n Sim(q,Dolce) = 0.5 * (1-1/30) * * 0 = 0.48 n Sim(q, Gabbana) = 0.5 (1-3/30) * * 1 = =

36 36

37 37

38 Query Tightening 38

39 39

40 [Ricci et al., 2002] 40

41 NutKing as a CBR System p p p p Problem = recommend a set of tourism related products and build a travel plan Cases = All the recommended travel plans that users have built using the system (how they were built and what they contain) Retrieval = search in the memory travel plans built during similar recommendation sessions Reuse 1. extract from previous travel plans elementary components (items) and use them to build a new plan 2. rank items found in the catalogues 41

42 Travel Plan Model and Interaction Session Collaborative Component 1: travel wish clf (family, bdg_medium,7,hotel) case Queries on content attributes (Golfing=True AND Nightlife=True) cnq (category=3 AND Health=True) Collaborative Component 2: selected products Travel bag rating (Kitzbühel,True,True,..) (Hotel Schwarzer, 3, True, ) 42

43 Item Ranking Input Current Case twc tb u r Q Suggest Q changes Travel components 1. Search the catalogue Interactive query management Locations from Catalogue loc1 loc2 loc3 4. Sort locations loc i by similarity to locations in reference cases Case Base 2. Search Similar Cases 3. Output Reference Set Case twc tb u r loc2 loc3 loc1 Ranked Items Output 43

44 Two-fold Similarity Sessions similarity Target user Target session s1 s2 s3 s4 s5 s6? Product similarity 44

45 Rank using Two-Fold Similarity p Given the current session case c and a set of retrieved products R (using the interactive query management facility - IQM) 1. retrieve 10 cases (c 1,, c 10 ) from the repository of stored cases (recommendation sessions managed by the system) that are most similar to c with respect to the collaborative features 2. extract products (p 1,, p 10 ) from cases (c 1,, c 10 ) of the same type as those in R 3. For each product r in R compute the Score(r) as the maximum of the product of a) the similarity of r with p i, b) the similarity of the current case c and the retrieved case c i containing p i 4. sort and display products in R according to the Score(r). 45

46 Example: Scoring Two Destinations Destinations matching the user s query D1 D2?? current case CC D1 current case CC D2 C1 CD1 C2 CD2 similar cases in the case base Score(D i ) = Max j {Sim(CC,C j )*Sim(D i,cd j )} Sim(CC,C1) 0.2 Sim(CC,C2) 0.6 Sim(D1, CD1) 0.4 Sim(D1, CD2) 0.7 Sim(D2, CD1) 0.5 Sim(D2, CD2) 0.3 Score(D1)=Max{0.2*0.4,0.6*0.7}=0.42 Score(D2)=Max{0.2*0.5,0.6*0.3}=

47 Tree-based Case Representation p A case is a rooted tree and each node has a: n node-type: similarity between two nodes in two cases is defined only for nodes with the same node-type n metric-type: node content structure - how to measure the node similarity with another node in a second case nt: cart nt: destination nt: case mt: vector mt: vector nt: destinations mt: set mt: vector nt: location mt: hierarchical cart1 dest1 X 1 c1 clf1 cnq1 dests1 accs1 acts1 dest2 ITEM X 2 X 3 X 4 47

48 Item Representation TRAVELDESTINATION=(X 1,X 2,X 3,X 4 ) Node Type Metric Type Example: Canazei X 1 LOCATION Set of hierarchical related symbols Country=ITALY, Region=TRENTINO, TouristArea=FASSA, Village=CANAZEI X 2 INTERESTS Array of Booleans Hiking=1, Trekking=1, Biking=1 X 3 ALTITUDE Numeric 1400 X 4 LOCTYPE Array of Booleans Urban=0, Mountain=1, Rivereside=0 X 1 = (Italy, Trentino, Fassa, Canazei) X 2 = (1,1,1) dest1 X 3 = 1400 X 4 = (0, 1, 0) 48

49 Item Query Language p For querying purposes items x a represented as simple vector features x=(x 1,, x n ) X 1 = (Italy, Trentino, Fassa, Canazei) X 2 = (1,1,1) dest1 X 3 = 1400 X 4 = (0, 1, 0) (Italy, Trentino, Fassa, Canazei, 1, 1, 1, 1400, 0, 1, 0) p A query is a conjunction of constraints over features: q=c 1 c 2 c m where m n and c k = x i k x i k l = true = v x i k u if x i if x if x k i i k k is boolean is nominal is numerical 49

50 Item Similarity dest1 If X and Y are two items with same node-type X 1 =(Italy, Trentino, Fassa, Canazei) X 2 = (1,1,1) X 3 = 1400 X 4 = (0, 1, 0) d(x,y) = (1/ i w i ) 1/2 [ i w i d i (X i,y i ) 2 ] 1/2 where 0 w i 1, and i=1..n (number of features). 1 if X i or Y i are unknown overlap(x i,y i ) if X i is symbolic X i - Y i /range i if X i is finite integer or real d i (X i,y i ) = Jaccard(X i,y i ) if X i is an array of Boolean Hierarchical(X i,y i ) if X i is a hierarchy Modulo(X i,y i ) if X i is a circular feature (month) Date (X i,y i ) if X i is a date Sim(X,Y) = 1 - d(x,y) or Sim(X,Y) = exp(- d(x,y)) 50

51 Item Similarity Example X 1 = (I, TN, Fassa, Canazei) Y 1 = (I, TN, Fassa,?) dest1 X 2 = (1,1,1) X 3 = 1400 Y 2 = (1,0,1) Y 3 = 1200 dest2 X 4 = (0, 1, 0) Y 4 = (1, 1, 0) Sim(dest 1, dest 2 ) = exp(!(1/ 4) d 1 (X 1,Y 1 ) 2 +!+ d 4 (X 4,Y 4 ) 2 ) = exp(!(1/ 4) (0.3) 2 + (1! 2 / 3) 2 + ((1400!1200) / 2000) 2 + (1!1/ 2) 2 ) = exp(!(1/ 4) 0, 461) = exp(!0,339) = 0, in the union 2 in the union 51

52 Case Distance nt: cart mt: vector nt: case nt: destinations mt: vector mt: set nt: destination mt: vector nt: location mt: hierarchical cart1 dest1 X 1 c1 clf1 cnq1 dests1 accs1 acts1 dest2 X 2 X 3 X 4 cart2 dest3 Y 1 c2 clf12 cnq2 dests2 accs2 acts2 dest4 dest5 Y 2 Y 3 Y 4 52

53 Case Distance 1 d( c cnq , c2) = W1d ( cart 3 1, cart 2) + W2d ( clf1, clf2) + W3d ( cnq1, 2) i= 1 W nt: case mt: vector cart1 i c1 clf1 cnq1 cart2 c2 clf12 cnq2 53

54 , cart2 ) = d( dests1, dests2 ) + d( accs 1, accs 2) d( act1, 2) d ( cart + act nt: cart nt: case mt: vector mt: vector cart1 c1 clf1 cnq1 dests1 accs1 acts1 cart2 c2 clf12 cnq2 dests2 accs2 acts2 54

55 55 c1 nt: case mt: vector clf1 cnq1 cart1 nt: cart mt: vector dests1 accs1 acts1 nt: destinations mt: set dest1 dest2 c2 clf12 cnq2 cart2 dests2 accs2 acts2 dest3 dest4 dest5 )), ( ), ( ), ( ), ( ), ( ), ( ( 2*3 1 ), ( dest dest d dest dest d dest dest d dest dest d dest dest d dest dest d dests dests d =

56 CBR Knowledge Containers p CBR is a knowledge-based approach to problem solving p The knowledge is contained into four containers n Cases: the instances belonging to our case base n Case representation language: the representation language that we decided to use to represent cases n Retrieval knowledge: the knowledge encoded in the similarity metric and in the retrieval algorithm n Adaptation knowledge: how to reuse a retrieved solution to solve the current problem. 56

57 Conclusions p Knowledge-based systems exploits knowledge to map a user to the products she likes p KB systems uses a variety of techniques p Knowledge-based systems requires a big effort in term of knowledge extraction, representation and system design p Many KB recommender systems are rooted in Case-Based Reasoning p Similarity of complex data objects is required often required in KB RSs. p NutKing is a hybrid case-based recommender system p The case is the recommendation session. 57

58 Questions p p p p p p p p p What are the main differences between a CF recommender system and a KB RS (such as activebuyers.com or Entree)? What is the role of query augmentation? What is the basic rationale of a CBR recommender system? What is a case in a CBR recommender system such as Entree? How a CBR recommender system learns to recommend? What are the knowledge containers is a CBR RS? What are the main differences between a classical CBR recommender system such as Entrée and Nutking? What are the motivations for the introduction of the doublesimilarity ranking method? What are the types of local similarity metrics used in Nutking? 58