An Analysis of Actors in Malay Films: Small Worlds, Centralities and Genre Diversity
Diana Purwitasari
Round 1
Reviewer 1 Report
The paper conducts a social network analysis by adding genre diversity to measure the quality and capacity of actors in the Malay language film industry. Measuring genre diversity uses the Normalized Herfindahl-Hirschman Index (NHHI).
Overall, the paper is good. Starting from writing to literature references that are adequate and convincing. However, there are a few things to note:
o Why only use data on 180 Malay language films?
o Does using the data of 180 Malay language films represent the entire Malay language film industry?
o Why does the paper only consider the relationship between actors?
o Why not consider the relationship between the director and actor? Having many collaborations between directors and actors also makes it possible to increase chemistry and film quality so that the chances of winning awards are greater.
It seems the same work has been published in Ref. 36.
Abstract: please elaborate more on the objective in this paper, i.e. the relations of network analysis measures with how to quantify the quality and capacity of actors. Why the mentions of TFA, DC, and so on will relate to the problem.
such as the sentence of
| Kevin Bacon has many films, and he is also diverse in his genre. |
Is there any scientific theory to support this statement with the current problem.
The research Design and Methods: it would be better to explain about hypothesis and then steps to prove it before showing the network analysis results as already describe in this paper.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
As far as I understand, the authors rely on the centrality metrics that are used for unweighted graphs. I think that it is a good starting point of research.
However, my main recommendation is to consider metrics that can be applied to weighted networks. By doing so, the the authors would be able to utilize the frequency of co-occurrence.
Further, I think that the authors can put more efforts on embedding their results in the existing body of literature. The single sentence "This in inline with past studies highlighting that 370 BC would measure different aspects compared to other centralities [22,23,38–41]." does not suffice.
Smaller comments:
1) The Erdos-Renyi model is more ancient than the Watts-Strogatz network.
2) When the authors compare the network of films against the Watts-Strogatz graph, they use the probability of rewiring p = <k> / (N-1).
I fully agree that small nonzero values of p guarantee that the underlying network will have quite large values of clustering and, at the same time, will feature a short average path length.
But can you explain the intuition behind this specific choice?
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors have given responses based on reviewer's comments.
Reviewer 2 Report
I thank the authors for revising the manuscript and I recommend it for publication
