Social Media Mining: An Introduction

P1: Trim: 6.125in×9.25in Top: 0.5in Gutter: 0.75in
CUUS2079-09 CUUS2079-Zafarani 978 1 107 01885 3 January 13, 2014 17:28

9.2 Classical Recommendation Algorithms 257

Most Pleasure.Unlike the least misery strategy, in the most pleasure

approach, we take the maximum rating in the group as the group rating:

Ri=maxu∈Gru,i. (9.35)

Since we recommend items that have the highestRivalues, this strategy

guarantees that the items that are being recommended to the group are

enjoyed the most by at least one member of the group.

Example 9.4.Consider the user-item matrix in Table9.3. Consider group

G={John,Jill,Juan}. For this group, the aggregated ratings for all prod-

ucts using average satisfaction, least misery, and maximum pleasure are as

follows.

Table 9.3.User-Item Matrix

Soda Water Tea Coffee

John 1 3 1 1

Joe 4 3 1 2

Jill 2 2 4 2

Jorge 1 1 3 5

Juan 3 3 4 5

Average Satisfaction:

RSoda=

1 + 2 + 3

3

= 2. (9.36)

RWater=

3 + 2 + 3

3

= 2. 66. (9.37)

RTea=

1 + 4 + 4

3

= 3. (9.38)

RCoffee=

1 + 2 + 5

3

= 2. 66. (9.39)

Least Misery:

RSoda=min{ 1 , 2 , 3 }= 1. (9.40)

RWater =min{ 3 , 2 , 3 }= 2. (9.41)

RTea=min{ 1 , 4 , 4 }= 1. (9.42)

RCoffee=min{ 1 , 2 , 5 }= 1. (9.43)