Social Network Features

Reason

Size/Degree

In mobile call log data, if an individual (ego) is having more social contacts then there is high chance that the individual (ego) will get the desired resources. Hence there is likely chance that an ego with higher degree may be productive.

Effective Size

We have used this feature as it is highly suitable with respect to our dataset, because a person with more effective size is considered to be having more unique social contacts and hence may be productive.

Betweenness Centrality

This is a popular feature in social network analysis and we have computed this feature from our mobile call log data. An actor with high betweenness centrality plays a powerful role in the network.

Density

We have used this popular feature in our analysis because dense networks are supposed to have more social capital and it is highly suitable for our dataset.

Tie Strength

We have used this feature because of its popularity and suitability with respect to our dataset where contact frequency and contact duration are indicators of tie strength.

Weak Ties

We have used this feature because of its popularity and importance with regard to our dataset.

Structural holes

We have used this feature because of its importance in ego networks. Ronald Burt has developed a number of measures to examine the position of each actor in their neighborhood.

K-Core

We have used this feature because it shows active participation of an individual in a social network if his/her at least K friends are also actively participating. This feature is intuitively appealing in our application.