http://www.mmds.

org

Introduction to Data Science

Yongfeng Zhang, Rutgers University
Supermarket shelf management – Market-basket
model:
 Goal: Identify items that are bought together by
sufficiently many customers
 Approach: Process the sales data collected with
barcode scanners to find dependencies among
items
 A classic rule:
▪ If someone buys diaper and milk, then he/she is
likely to buy beer
▪ Don’t be surprised if you find six-packs next to diapers!
2
 Input:
 A large set of items TID Items
▪ e.g., things sold in a 1 Bread, Coke, Milk
supermarket 2 Beer, Bread
3 Beer, Coke, Diaper, Milk
 A large set of baskets 4 Beer, Bread, Diaper, Milk
 Each basket is a 5 Coke, Diaper, Milk

small subset of items

Output:
▪ e.g., the things one
customer buys on one day Rules Discovered:
{Milk} --> {Coke}
 Want to discover {Diaper, Milk} --> {Beer}
association rules
▪ People who bought {x,y,z} tend to buy {v,w}
▪ Amazon! (“Frequently bought together” recommendation)

3
 Items = products; Baskets = sets of products
someone bought in one trip to the store
 Real market baskets: Chain stores keep TBs of
data about what customers buy together
▪ Tells how typical customers navigate stores, lets
them position tempting items
▪ Suggests tie-in “tricks”, e.g., run sale on diapers
and raise the price of beer
▪ Need the rule to occur frequently, or no $$’s
 Amazon’s people who bought X also bought Y
4
 Baskets = sentences; Items = documents
containing those sentences
▪ Items that appear together too often could represent
plagiarism
▪ Notice items do not have to be “in” baskets
▪ You only need to know each item belongs to which basket(s)

 Baskets = patients; Items = drugs & side-effects

▪ Has been used to detect combinations
of drugs that result in particular side-effects
▪ But requires extension: Absence of an item
needs to be observed as well as presence
5
 A general many-to-many mapping
(association) between two kinds of things
▪ But we ask about connections among “items”,
not “baskets”

 For example:
▪ Finding communities in graphs (e.g., Twitter)

6
 Finding communities in graphs (e.g., Twitter)
 Baskets = nodes; Items = outgoing neighbors
▪ Searching for complete bipartite subgraphs Ks,t of a
big graph
 How?
▪ View each node i as a
basket Bi of nodes i it points to
t nodes
s nodes

▪ Ks,t = a set Y of size t that

occurs in s buckets Bi
…

▪ Looking for Ks,t → all frequent

A dense 2-layer graph sets of size t with support s

7
First: Define
Frequent itemsets
Association rules:
Confidence, Support, Interestingness
Then: Algorithms for finding frequent itemsets
Finding frequent pairs
A-Priori algorithm
PCY algorithm + 2 refinements
8
 Simplest question: Find sets of items that
appear together “frequently” in baskets
 Support for itemset I: Number of baskets
containing all items in I TID Items

▪ (Often expressed as a fraction 1

2
Bread, Coke, Milk
Beer, Bread

of the total number of baskets) 3

4
Beer, Coke, Diaper, Milk
Beer, Bread, Diaper, Milk

 Given a support threshold s, 5 Coke, Diaper, Milk

Support of
then sets of items that appear {Beer, Bread} = 2
in at least s baskets are called
frequent itemsets
9
 Items = {milk, coke, pepsi, beer, juice}
 Support threshold = 3 baskets
B1 = {m, c, b} B2 = {m, p, j}
B3 = {m, b} B4 = {c, j}
B5 = {m, p, b} B6 = {m, c, b, j}
B7 = {c, b, j} B8 = {b, c}
 Frequent itemsets: {m}, {c}, {b}, {j},
{m,b} , {b,c} , {c,j}.

10
 Association Rules:
If-then rules about the contents of baskets
 {i1, i2,…,ik} → j means: “if a basket contains
all of i1,…,ik then it is likely to contain j”
 In practice there are many rules, want to find
significant/interesting ones!
 Confidence of this association rule is the
probability of j given I = {i1,…,ik}
support( I  j )
conf( I → j ) =
support( I )
11
 Not all high-confidence rules are interesting
▪ The rule X → milk may have high confidence for many
itemsets X, because milk is just purchased very often
(independent of X) and the confidence will be high
 Interest of an association rule I → j:
difference between its confidence and the
fraction of baskets that contain j
Interest(I → j ) = conf( I → j ) − Pr[ j ]
▪ Interesting rules are those with high positive or
negative interest values (usually above 0.5)
Positive: high conf low Pr, j is rare but it frequently co-occurs with I (complementary)
Negative: low conf high Pr, j is frequent but seldom co-occurs with I (substitutive)

12
 B1 = {m, c, b} B2 = {m, p, j}
B3 = {m, b} B4= {c, j}
B5 = {m, p, b} B6 = {m, c, b, j}
B7 = {c, b, j} B8 = {b, c}

 Association rule: {m, b} →c

▪ Confidence = 2/4 = 0.5
▪ Interest = |0.5 – 5/8| = 1/8
▪ Item c appears in 5/8 of the baskets
▪ Rule is not very interesting!

13
 Problem: Find all association rules with
support ≥s and confidence ≥c
▪ Note: Support of an association rule is the support
of the set of items on the left side
 Hard part: Finding the frequent itemsets!
▪ If {i1, i2,…, ik} → j has high support and
confidence, then both {i1, i2,…, ik} and
{i1, i2,…,ik, j} will be “frequent”

support( I  j )
conf( I → j ) =
support( I )
14
 Step 1: Find all frequent itemsets I
▪ (we will explain this next)
 Step 2: Rule generation
▪ For every subset A of I, generate a rule A → I \ A
▪ Since I is frequent, A is also frequent
▪ Variant 1: Single pass to compute the rule confidence
▪ confidence(A,B→C,D) = support(A,B,C,D) / support(A,B)
▪ Variant 2:
▪ Observation: If A,B,C→D is below confidence, so is A,B→C,D
▪ Can only generate “bigger” rules from smaller ones!
▪ Output the rules above the confidence threshold
15
 B1 = {m, c, b} B2 = {m, p, j}
B3 = {m, c, b, n} B4= {c, j}
B5 = {m, p, b} B6 = {m, c, b, j}
B7 = {c, b, j} B8 = {b, c}
 Support threshold s = 3, confidence c = 0.75
 1) Frequent itemsets:
▪ {b,m} {b,c} {c,m} {c,j} {m,c,b}
 2) Generate rules:
▪ b→m: c=4/6 b→c: c=5/6 b,c→m: c=3/5
▪ m→b: c=4/5 … b,m→c: c=3/4
▪ b→c,m: c=3/6
…
16
 To reduce the number of rules we can
post-process them and only output:
▪ Maximal frequent itemsets: e.g., “{A, B, C} is frequent”
No superset is frequent is more informative than
“{A, B}” is frequent
▪ i.e., All supersets are infrequent
▪ Gives more pruning
or
▪ Closed itemsets:
No superset has the same count
▪ i.e., All supersets are less frequent
▪ But supersets can still be frequent
▪ Stores not only frequent information, but exact counts
17
Support threshold = 3
Frequent, but
superset BC
Support Maximal(s=3) Closed also frequent.

A 4 No No Frequent, and
B 5 No Yes its only superset,
ABC, not freq.
C 3 No No Superset BC
AB 4 Yes Yes has same count.

AC 2 No No Its only super-

set, ABC, has
BC 3 Yes Yes smaller count.

ABC 2 No Yes

18
  Back to finding frequent itemsets Item
Item

 Typically, data is kept in flat files Item

Item

rather than in a database system: Item

Item

▪ Stored on disk Item

Item

▪ Stored basket-by-basket Item

Item

▪ Baskets are small but we have Item

Item

many baskets and many items

▪ Expand baskets into pairs, triples, etc. Etc.
as you read baskets
▪ Use k nested loops to generate all
sets of size k
Items are positive integers,
Note: We want to find frequent itemsets. To find them, we and boundaries between
have to count them. To count them, we have to generate them. baskets are –1.
20
 The hardest problem often turns out to be
finding the frequent pairs of items {i1, i2}
▪ Why? Freq. pairs are common, freq. triples are rare
▪ Why? Probability of being frequent drops exponentially
with size; number of sets grows more slowly with size

 Let’s first concentrate on pairs, then extend to

larger sets

21
 Naïve approach to finding frequent pairs
 Read file once, counting in main memory
the occurrences of each pair:
▪ From each basket of n items, generate its
n(n-1)/2 pairs by two nested loops

 E.g., Illustrate the process by counting the

pairs with two baskets {A, B, C} and {A, B, D}

22
 Naïve approach to finding frequent pairs
 Read file once, counting in main memory
the occurrences of each pair:
▪ From each basket of n items, generate its
n(n-1)/2 pairs by two nested loops
 Fails if O(#items2) exceeds main memory
▪ Remember: #items can be
100K (Wal-Mart) or 10B (Web pages)
▪ Suppose 105 items, counts are 4-byte integers
▪ Number of pairs of items: 105(105-1)/2 = 5*109
▪ Therefore, 2*1010 (20 gigabytes) of memory needed
23
 For many frequent-itemset algorithms,
main-memory is the critical resource
▪ As we read baskets, we need to count
something, e.g., occurrences of pairs of items
▪ The number of different things we can count
is limited by main memory
▪ Swapping counts in/out is a disaster
▪ Randomly reading/writing disk is very time consuming

A naïve memory-based algorithm would not work

24
Two approaches:
 Approach 1: Count all pairs using a matrix
 Approach 2: Keep a table of triples [i, j, c] =
“the count of the pair of items {i, j} is c.”
▪ If integers and item ids are 4 bytes, we need
approximately 12 bytes for pairs with count > 0
▪ Plus some additional overhead for the hashtable
Note:
 Approach 1 only requires 4 bytes per pair
 Approach 2 uses 12 bytes per pair
(but only for pairs with count > 0)
25
 12 per
4 bytes per pair
occurring pair

Triangular Matrix Triples

26
 Approach 1: Triangular Matrix
▪ n = total number items
▪ Count pair of items {i, j} only if i<j
▪ Keep pair counts in lexicographic order:
▪ {1,2}, {1,3},…, {1,n}, {2,3}, {2,4},…,{2,n}, {3,4},…
▪ Pair {i, j} is at position (i –1)(n– i/2) + j –1
▪ Total number of pairs n(n –1)/2; total bytes= 2n(n-1)
▪ Triangular Matrix requires 4 bytes per pair
 Approach 2 uses 12 bytes per occurring pair
(but only for pairs with count > 0)
▪ Beats Approach 1 if less than 1/3 of
possible pairs actually occur
27
 Approach 1: Triangular Matrix
▪ n = total number items
▪ Count pair of items {i, j} only if i<j
▪ Keep pair counts in lexicographic order:
Problem is if we have too
▪ {1,2}, {1,3},…, {1,n}, {2,3}, {2,4},…,{2,n}, {3,4},…
▪ Pair {i,many items(iso
j} is at position thei/2)pairs
–1)(n– + j –1
▪ Total number
do notoffit pairs n(n –1)/2;
into total bytes= 2n(n-1)
memory.
▪ Triangular Matrix requires 4 bytes per pair
 ApproachCan we
2 uses 12 do better?
bytes per pair
(but only for pairs with count > 0)
▪ Beats Approach 1 if less than 1/3 of
possible pairs actually occur
28
 In practice, association-rule algorithms read Item
Item
the data in passes – all baskets read in turn Item
Item
Item
Item
 Algorithm may read the data for multiple passes Item
Item
Item
Item
 The true cost of mining massive disk-resident Item
Item
data is usually the number of disk I/Os
 We measure the cost by the number of Etc.

passes an algorithm makes over the data

30
 A two-pass approach called
A-Priori limits the need for
main memory
 Key idea: monotonicity
▪ If a set of items I appears at
least s times, so does every subset J of I
 Contrapositive for pairs:
If item i does not appear in s baskets, then no
pair including i can appear in s baskets
 So, how does A-Priori find frequent pairs?
31
 Pass 1: Read baskets and count in main memory
the occurrences of each individual item
▪ Requires only memory proportional to #items

 Items that appear ≥ 𝒔 times are the frequent items

 Pass 2: Read baskets again and count in main
memory only those pairs where both elements
are frequent (from Pass 1)
▪ Requires memory proportional to O(#frequent-items2)
only (for counts)
▪ Plus a list of the frequent items (so you know what must
be counted)
32
 Frequent items
Item counts
Main memory

Counts of
pairs of
frequent items
(candidate
pairs)

Pass 1 Pass 2

33
 For each k, we construct two sets of
k-tuples (sets of size k):
▪ Ck = candidate k-tuples = those that might be
frequent k-tuple sets (support > s)
▪ based on information from the pass for k–1
▪ Lk = the set of truly frequent k-tuples
Count All pairs Count
All of items To be
the items the pairs explained
items from L1

C1 Filter L1 Construct C2 Filter L2 Construct C3

35
 ** Note here we generate new candidates by
generating Ck from Lk-1 and L1.
But that one can be more careful with candidate
generation. For example, in C3 we know {b,m,j}
cannot be frequent since {m,j} is not frequent

 Steps of the A-Priori algorithm

▪ C1 = { {b} {c} {j} {m} {n} {p} }
▪ Count the support of itemsets in C1
▪ Prune non-frequent: L1 = { b, c, j, m }
▪ Generate C2 = { {b,c} {b,j} {b,m} {c,j} {c,m} {j,m} }
▪ Count the support of itemsets in C2
▪ Prune non-frequent: L2 = { {b,m} {b,c} {c,m} {c,j} }
▪ Generate C3 = { {b,c,m} {b,c,j} {b,m,j} {c,m,j} }
▪ Count the support of itemsets in C3 **

▪ Prune non-frequent: L3 = { {b,c,m} }

36
 One pass for each k (itemset size)
 Needs room in main memory to count
each candidate k–tuple
 For typical market-basket data and reasonable
support (e.g., 1%), k = 2 requires the most memory
 Many possible extensions:
▪ Association rules with constraints:
▪ For example: Users over 65 have 2 cars
▪ Association rules when items are in a taxonomy
▪ Bread, Butter → AppleJam, StrawberryJam
▪ Bread, Butter → FruitJam
▪ Dynamically lower the support s as itemset gets bigger
37
 Observation:
In pass 1 of A-Priori, most memory is idle
▪ We store only individual item counts
▪ Can we use the idle memory to reduce
memory required in pass 2?
Item counts Frequent items
Main memory

Counts of
pairs of
frequent items
(candidate
pairs)

Pass 1 Pass 2 39
 Observation:
In pass 1 of A-Priori, most memory is idle
▪ We store only individual item counts
▪ Can we use the idle memory to reduce
memory required in pass 2?
 Pass 1 of PCY: In addition to item counts,
maintain a hash table with as many
buckets as fit in memory
▪ Hash function should be deterministic
▪ Keep a count for each bucket into which
pairs of items are hashed
▪ For each bucket just keep the count, not the actual
pairs that hash to the bucket!
40
 FOR (each basket) :
FOR (each item in the basket) :
add 1 to item’s count;
New FOR (each pair of items) :
in hash the pair to a bucket;
PCY add 1 to the count for that bucket;

 Few things to note:

▪ Pairs of items need to be generated from the input
file; they are not present in the file
▪ We are not just interested in the presence of a pair,
but we need to see whether it is present at least s
(support) times
41
 Observation: If a bucket contains a frequent pair,
then the bucket is surely frequent
 However, even without any frequent pair,
a bucket can still be frequent 
▪ So, we cannot use the hash value to eliminate any
member (pair) of a “frequent” bucket
 But, for a bucket with total count less than s,
none of its pairs can be frequent ☺
▪ Pairs that hash to this bucket can be eliminated as
candidates

 Pass 2:
Only count pairs that hash to frequent buckets
42
 How do we reserve memory for pass-2?
 Replace the buckets by a bit-vector:
▪ 1 means the bucket count exceeded the support s
(call it a frequent bucket); 0 means it did not

 4-byte integer counts are replaced by bits,

so the bit-vector requires 1/32 of memory
▪ 1 byte = 8 bits Item counts

Hash
 Also, decide which items are frequent Hash table
table
for pairs
and list them for the second pass
43
 Item counts Frequent items

Bitmap
Main memory

Hash
Hash table
table Counts of
for pairs
candidate
pairs

Pass 1 Pass 2

44
 Count all pairs {i, j} that meet the
conditions for being a candidate pair:
1. Both i and j are frequent items
2. The pair {i, j} hashes to a bucket whose bit in
the bit vector is 1 (i.e., a frequent bucket)

 Both conditions are necessary for the

pair to have a chance of being frequent

45
 Key idea: After Pass 1 of PCY, rehash only
those pairs that qualify for Pass 2 of PCY
▪ i and j are frequent, and
▪ {i, j} hashes to a frequent bucket from Pass 1

 In the middle pass, we can filter out even

more infrequent pairs

 Requires 3 passes over the data

47
 Item counts Freq. items Freq. items
Main memory

Bitmap 1 Bitmap 1

First Bitmap 2
hash table
First
Second Counts
hash table Counts ofof
hash table candidate
candidate
pairs
pairs

Pass 1 Pass 2 Pass 3

Hash pairs {i,j} Count pairs {i,j} iff:
Count items into Hash2 iff: i,j are frequent,
Hash pairs {i,j} i,j are frequent, {i,j} hashes to
{i,j} hashes to freq. bucket in B1
freq. bucket in B1 {i,j} hashes to
freq. bucket in B2
48
 Count only those pairs {i, j} that satisfy these
candidate pair conditions:
1. Both i and j are frequent items
2. Using the first hash function, the pair hashes to
a bucket whose bit in the first bit-vector is 1
3. Using the second hash function, the pair hashes to
a bucket whose bit in the second bit-vector is 1

49
1. The two hash functions have to be
independent
2. We need to check both hashes on the
third pass
▪ If not, we may count pairs of frequent items
i and j that is first hashed to a frequent
bucket but happened to hash to an
infrequent bucket in the second hash

50
 Key idea: Use several independent hash
tables on the first pass

Item counts Freq. items

Bitmap 1
Main memory

First
First hash
hash table
table Bitmap 2

Counts
Countsofof
Second
Second candidate
candidate
hash table
hash table pairs
pairs

Pass 1 Pass 2

51
 Key idea: Use several independent hash
tables on the first pass

 Risk: Halving the number of buckets doubles

the average count
▪ We have to be sure most buckets will still not
reach count s

 If so, we can get a benefit like multistage,

but in only 2 passes

52
 Either multistage or multihash can use more
than two hash functions
 In multistage, there is a point of diminishing
returns, since the bit-vectors eventually
consume all of main memory

 For multihash, the bit-vectors occupy exactly

what one PCY bitmap does, but too many
hash functions makes most counts > s

53
 A-Priori, PCY, etc., take k passes to find
frequent itemsets of size k
 Can we use fewer passes?
 Use 2 or fewer passes for all sizes,
but may miss some frequent itemsets
▪ Random sampling
▪ SON (Savasere, Omiecinski, and Navathe)
▪ Toivonen (see textbook)

55
 Take a random sample of the market baskets
 Run a-priori or one of its improvements
in main memory Copy of
▪ So we don’t pay for disk I/O each sample

Main memory
baskets
time we increase the size of itemsets
▪ Reduce support threshold
Space
proportionally to match the sample size for
counts
▪ e.g. sample 10% of the baskets, then use s/10

56
 Optionally, verify that the candidate pairs are
truly frequent in the entire data set by a
second pass (avoid false positives)

 But you don’t catch sets that are frequent in

the whole but not in the sample
▪ Smaller threshold, e.g., s/125, helps catch more
truly frequent itemsets
▪ But requires more space

57
 Repeatedly read small subsets of the baskets
into main memory and run an in-memory
algorithm to find all frequent itemsets
▪ Note: we are not sampling, but processing the
entire file in memory-sized chunks

 An itemset becomes a candidate if it is found

to be frequent in any one or more subsets of
the baskets.

58
 On a second pass, count all the candidate
itemsets and determine which are frequent in
the entire set

 Assumption: an itemset cannot be frequent in

the entire set of baskets unless it is frequent
in at least one subset.

 Also, may miss some itemsets that are

frequent in entire set but not in any subset.
▪ Trade-off between efficiency and accuracy
59
 SON lends itself to distributed data mining
 Baskets distributed among many nodes
▪ Compute frequent itemsets at each node
▪ Distribute candidates to all nodes
▪ Accumulate the counts of all candidates

60