web science-2019讨论区

Group 13:Discussion about session 2

 
Picture of kx m
Group 13:Discussion about session 2
by kx m - Tuesday, 21 May 2019, 9:49 PM
 

The students from group 13 will discuss about session 1 in this page.

Picture of kx m
回复: Group 13:Discussion about session 2
by kx m - Tuesday, 21 May 2019, 9:51 PM
 

Let's begin with the third question. Do you think these analyses are a good way of measuring power and influence in a social group? What might they miss?

Picture of 航 薛
回复: Group 13:Discussion about session 2
by 航 薛 - Tuesday, 21 May 2019, 10:04 PM
 

I think the three indicators mentioned in video are relatively comprehensive for the measurement of power, but it is difficult to determine their weights. And these three indicators are difficult to measure and quantify by the same standard.

Picture of 航 薛
回复: Group 13:Discussion about session 2
by 航 薛 - Tuesday, 21 May 2019, 10:11 PM
 

For 'what might they miss?'. Imagine this scenario. Maybe A is a leader, so he has a lot of power. But when he issues orders, he may only send decisions to a few people, and the decisions will be divided step by step, that is, there will be no RT.Meanwhile the corresponding replies may be relatively few. Only the high-level who could contact with A will reply him or her.

But I haven't figured out exactly how to do that with an index.

Picture of kx m
回复: Group 13:Discussion about session 2
by kx m - Tuesday, 21 May 2019, 10:36 PM
 

I agree with you. In my opinion, we should not only consider the number of conections, but also the direction of information flow. For example, the power and influence of the boss is also reflected in the orders he issued will be passed on to everyone.

Picture of 添瑞 郭
回复: Group 13:Discussion about session 2
by 添瑞 郭 - Tuesday, 21 May 2019, 9:55 PM
 

In my view, the way mentioned in video is an efficient way to measure power and influence of one people, the different methods bring to different results also shows its comprehensiveness. But i think it miss the width of the links between different nodes which can represent the intimate degree between two persons. I think the news delivered by friends is faster than acquaintance.

Picture of Hengyi Lin
回复: Group 13:Discussion about session 2
by Hengyi Lin - Tuesday, 21 May 2019, 10:19 PM
 

I agree with you. Actually the most part of  reply or retwits is the interaction between friends, and the more science way to measure the power is to combine these method (because they're all resonable and effective method) up. One method only concern about one aspect of the problem, but we can give weight value to result of every method then make a comprehensive decision and avoid the shortsighted judgement.

Picture of 小乐 段
回复: Group 13:Discussion about session 2
by 小乐 段 - Thursday, 23 May 2019, 4:26 AM
 

I agree with that, but the hardest part of talking about it is quantifying, what criteria to use to weight it, how much to refine it, and that's a lot of discussion.

Picture of 小乐 段
回复: Group 13:Discussion about session 2
by 小乐 段 - Thursday, 23 May 2019, 4:31 AM
 

I wonder if we can simulate it with a small interactive network starting from 2. We first identify all the indicators and details that we want to consider and try to simplify them as much as possible. These indicators are then quantified. Try to compute it, or just write out the algorithm. At last, a simple generalization is carried out and more complicated logical relations are discarded.