Locating the Epidemic Source in Complex Networks with Sparse Observers
Round 1
Reviewer 1 Report
The paper is well written, clear and with a strong contribution in the field. I recommand this paper for publication in Applied Sciences. I just formulate a very minor remark: a paper organization will improve the presentation for the reader.
Author Response
Thank you for your excellent suggestion!
I have checked the paper seriously and corrected some impertinent points.
Reviewer 2 Report
The English has to be revised. There is some errors done repeatedly in the text. Some sentences are not complete. Line 116: It should be θvu instead of α. Line 158: Since t0 is an integer, its range should be {t0l, t0l+1, ..., t0r}. Line 172: Can you provide the value of equation (10) for the other possible values of t0. Line 198: ERT should stand for Erdös-Rényi tree Section 3.3: Since the performance of the PESL algortihm depends heavily on the values of the parameters, how someone can use it in practice? Can you suggest (or refer to) a method to estimate them. Line 273: The data were taken on June 23, 2013 and or to July 3, 2013. Figure 7: How do the values of γ and δ are chosen? Can you add a reference or a justification? Reference 1: The title of this article is missing. Reference 28: This is the same as Reference 6.Author Response
Thank you for your excellent suggestions!
I have already read the suggestions you reviewed seriously. It can be reduced to six text errors, one supplement and two questions.
Firstly, I have collected the six text errors point by point in line 116, line 158, line 198, line 273, reference 1 and reference 28.
Secondly, I answered the question and provided one supplement: Can you provide the value of equation (10) for the other possible values of t0
We provided the value of equation (10) for the other possible values of t0, when t0=-2, L=-7.18; when t0=-1, L=-4.84.
Finally, I will next answer the following two questions.
1.Section 3.3: Since the performance of the PESL algortihm depends heavily on the values of the parameters, how someone can use it in practice? Can you suggest (or refer to) a method to estimate them.
The parameters which affect the performance of the PESL algorithm are infection probability β and recovery probability γ. These two parameters are acquired by a lot of statistics when the epidemic outbroke and meanwhile it refers to the data of similar epidemic. Because a susceptible one is infected with probability 1-(1-β)n when it has n infectious neighbors, we get the conclusion--the actual infection probability from t to t+1 equals the average value of all 1-(1-β)n at time t so that we can estimate the infection probability β. For recovery probability γ, we just calculate the probability where the infectious one recovered to recovery state at time t.
2. Figure 7: How do the values of γ and δ are chosen? Can you add a reference or a justification?
Figure 7 shows the relationship between the performance of source localization and infection probability β, so we need to keep the recovery probability γ and reporting probability δ invariable. Our study applied in highly lethal epidemic is of great significance, so we make the γ=0.1. When a new epidemic outbroke, only partial infectious people recognized it as epidemic without chosing to go to the hospital or the medical equipment is simple and the medical staff is poor in some area, which means the reporting probability is small. δ=0.3 is rensonable. These are all reasonable assumption.
