Assignment 7 (Sol.

)
Introduction to Data Analytics
Prof. Nandan Sudarsanam & Prof. B. Ravindran

1. Let X, Y be two itemsets, and let supp(X) denote the support of itemset X. Then the
confidence of the rule X → Y , denoted by conf (X → Y ) is
supp(X)
(a) supp(Y )
supp(Y )
(b) supp(X)
supp(X∪Y )
(c) supp(X)
supp(X∪Y )
(d) supp(Y )
supp(X∩Y )
(e) supp(X)

Sol. (c)
Confidence measures the probability of seeing items in the consequent (RHS) of the rule given
that we have observed items in the antecedent (LHS) of the rule in a transaction.
2. In identifying frequent itemsets in a transactional database, we find the following to be the
frequent 3-itemsets: {B, D, E}, {C, E, F}, {B, C, D}, {A, B, E}, {D, E, F}, {A, C, F}, {A,
C, E}, {A, B, C}, {A, C, D}, {C, D, E}, {C, D, F}, {A, D, E}. Which among the following
4-itemsets can possibly be frequent?

(a) {A, B, C, D}
(b) {A, B, D, E}
(c) {A, C, E, F}
(d) {C, D, E, F}

Sol. (d)
By the apriori property, only itemset {C, D, E, F} can possibly be frequent since all of its
subsets of size 3 are listed as frequent. The other 4-itemsets cannot be frequent since not all
of their subsets of size 3 are frequent. For example, for the first option, the itemset {A, B, D}
is not frequent.
3. Let X, Y be two itemsets, supp(X) denote the support of itemset X and conf (X → Y ) denote
the confidence of the rule X → Y , denoted by conf (X → Y ). Then lift of the rule, denoted
by lif t(x → Y is
supp(X)
(a) supp(Y )

1
 supp(X)×supp(Y )
(b) supp(Y )
supp(X∪Y )
(c) supp(X)
supp(X∪Y )
(d) supp(X)×supp(Y )
supp(X∩Y )
(e) supp(X)×supp(Y )

Sol. (d)
The lift of a rule can be thought of as the ratio of the observed support to the support that
would be expected if the antecedent and consequent were independent.
4. Consider the following transactional data.

Transaction ID Items
1 A, B, E
2 B, D
3 B, C
4 A, B, D
5 A, C
6 B, C
7 A, C
8 A, B, C, E
9 A, B, C

Assuming that the minimum support is 2, what is the number of frequent 2-itemsets (i.e.,
frequent items sets of size 2)?

(a) 2
(b) 4
(c) 6
(d) 8

Sol. (c)
Candidate 1-itemsets:
itemset support
{A} 6
{B} 7
{C} 6
{D} 2
{E} 2

Frequent 1-itemsets:

itemset support
{A} 6
{B} 7
{C} 6
{D} 2
{E} 2

2
 Candidate 2-itemsets:

itemset support
{A, B} 4
{A, C} 4
{A, D} 1
{A, E} 2
{B, C} 4
{B, D} 2
{B, E} 2
{C, D} 0
{C, E} 1
{D, E} 0

Frequent 2-itemsets:

itemset support
{A, B} 4
{A, C} 4
{A, E} 2
{B, C} 4
{B, D} 2
{B, E} 2

5. For the same data as above, what are the number of candidate 3-itemsets and frequent 3-
itemsets respectively?

(a) 1, 1
(b) 2, 2
(c) 2, 1
(d) 3. 2

Sol. (b)
Candidate 3-itemsets:

itemset support
{A, B, C} 2
{A, B, E} 2

Frequent 3-itemsets:

itemset support
{A, B, C} 2
{A, B, E} 2

6. Continuing with the same data, how many association rules can be derived from the frequent
itemset {A, B, E}? (Note: for a frequent itemset X, consider only rules of the form S -¿ (X-S),
where S is a non-empty subset of X.)

3
 (a) 3
(b) 6
(c) 7
(d) 8

Sol. (b)
{A} → {B, E}
{B} → {A, E}
{E} → {A, B}
{A, B} → {E}
{A, E} → {B}
{B, E} → {A}
7. For the same frequent itemset as mentioned above, which among the following rules have a
minimum confidence of 60%?

(a) A ∧ B =⇒ E
(b) A ∧ E =⇒ B
(c) E =⇒ A ∧ B
(d) A =⇒ B ∧ E

Sol. (b), (c)

The confidence values for the above four rules are respectively, 2/4, 2/2, 2/2, and 2/6. Hence,
only rules in (b) and (c) have the minimum required confidence.
8. Suppose we are given a large text document and the aim is to count the words of different
lengths, i.e., our output will be of the form - x words of length 1, y words of length 2, and
so on. Assuming a map-reduce approach to solving this problem, which among the following
key-value outputs would you prefer for the map phase? (Hint: consider the solution for the
reduce part asked in the next question as well to ensure a complete algorithm to solve the
problem.)

(a) key - word, value - length (of corresponding word)

(b) key - word, value - 1
(c) key - length (of corresponding word), value - word
(d) key - 1, value - word

Sol. (c)
Given a word, in the map phase we create a key-value pair, where the key is the length of the
word and the value is the word itself.
9. For the above question, what would be the appropriate processing action in the reduce phase?

(a) for each key which is a word, compute the sum of the values corresponding to this key
(b) for each key which is a number, compute the lengths of the words in the corresponding
list of values

4
 (c) for each key which is a number, count the number of words in the corresponding list of
values

Sol. (c)
In the reduce phase, all words of the same size will be available in the same reduce node. Thus,
in each reduce node, counting the number of words in the list of values will give us the number
of words of a particular length observed in the original document.
10. Let d1 and d2 be two distances according to some distance measure d. A function f is said to
be (d1 , d2 , p1 , p2 )-sensitive if

(a) if d(x, y) ≤ d1 , then the probability that f (x) = f (y) is at least p1

(b) if d(x, y) ≥ d2 , then the probability that f (x) = f (y) is at most p2
where d(·, ·) is a distance function. Given such a (d1 , d2 , p1 , p2 )-sensitive function, a better
function (for use in locality sensitive hashing) would be one with

(a) an increased value of p1

(b) a decreased value of p1
(c) an increased the value of p2
(d) a decreased the value of p2

Sol. (a), (d)

Compared to the original function, if we can increase the value of p1 , we can ensure that if two
points are close enough (d(x, y) ≤ d1 ), then the probability of a collision is higher. This is a
desirable property when performing locality sensitive hashing. Similarly, if p2 can be reduced
it would indicate that given some separation between two points (d(x, y) ≥ d2 ), the probability
of still observing a collision (which is undesirable) is reduced.