Political Signed Temporal Networks: A Deep Learning Approach
Round 1
Reviewer 1 Report
Paper deals with important topics Politics. The authors have developed a prototype of the evolution of political signed networks. However, I have a number of suggestions: 1. Abstract should be extended by the obtained results in part of the performance evaluating the DL-based approach. 2. In the section Proposed Approach authors should analyze different Deep Learning Algorithms and provide a cleaner idea, in which cases your approach performs better. 3. Authors should clearly point-by-point describe the main contributions of this paper. It should somehow resonate with the title of the work.4. Proposed Deep Leaning data flow unclear. I would suggest providing a better visual description and architecture approach presentation. 5. In addition to the accuracy indicators, it would be good to see the time of the training procedure for all investigated deep learning methods or to note this point in the perspective of further research. Please provide an accuracy rate, as you mentioned error close to zero. 6. Conclusion section should be extended by the obtained from the authors' results.
7. A lot of references are outdated and unlinked. Please fix it by using 3-5 years old papers in high-impact journals.
Author Response
Dear Sir,
Attached you´ll find the revisions and/or rebuttal according to each of the comments and suggestions proposed:
Point 1: Please see lines 5-10
Point 2: From our point of view the comparisons with other deep learning algorithms is not necessary since the results would vary minimally, here we are interested in generalization performance and to this respect the architecture selection is more important rather than the training algorithm . Furthermore, most of the deep learning algorithms are gradient-based: clipped gradients, Truncated gradients and so forth. The scaled conjugate gradients is a top performance learning algorithm that solved the vanishing gradients problems and is commonly used because of its convergence properties and acceptable computational complexity compared, for example with quasi-Newton methods or the problems associated with standard algorithms like BackPropagation.
Point 3: Please see lines 422-438
Point 4: A new figure (figure 1) has been created and inserted in the text giving a schematic vision of the data flow.
Point 5: Please see lines 362-394
Point 6: Please see lines 639-646
Point 7: The following references has been inserted (mainly in the introduction):
\bibitem[Elhesha(2019)]{Elhesha}
Elhesha R., Sarkar A., Boucher J., \& Kahveci T. (2019) Identification of co-evolving temporal networks , \emph{BMC Genomics}, Vol. 20(Suppl 6):434, doi: 10.1186/s12864-019-5719-9. https://pubmed.ncbi.nlm.nih.gov/31189471/
\bibitem[Wang \& Gan(2021)]{Wang}
Wang W., Wuang L., \& Gan C. (2021). Epidemic Spreading Dynamics on Temporal Networks, Collection of papers 2021, https://www.hindawi.com/journals/complexity/si/813208/
\bibitem[Masuda(2017)]{Masuda}
Masuda N. , \& Holme P. (Eds.) (2017). \emph{Temporal Network Epidemiology (Theoretical Biology)} 1st Edition, Springer, https://link.springer.com/book/10.1007/978-981-10-5287-3.
\bibitem[Zhu et al.(2022)]{Zhu} Zhu, Z., Sun Y., Kuang Y., Yuan X., Gu H., Zhu J., \& Xing W. (2022). Contemporaneous symptom networks of multidimensional symptom experiences in cancer survivors: A network analysis. \emph{Cancer Medicine} Wiley, pp. 1-11. https://doi.org/10.1002/cam4.4904.
\bibitem[Jordan et al.(2020)]{Jordan} Jordan G., Winer S., \& Salem T. The current status of temporal network analysis for clinical science: Considerations as the paradigm shifts?. https:/onlinelibrary.wiley.com/doi/10.1002/jclp.22957.
\bibitem[Nasuha-Daud et al.(2020)]{Nasuha-Daud} Nasuha-Daud N., Hafizah-Ab-Hamid S., Saadoon M., Sahran F., \& Badrul-Anuar N.(2020). Applications of link prediction in social networks: A review, \emph{Journal of Network and Computer Applications}, Vol. 166, 2020, 102716. https://doi.org/10.1016/j.jnca.2020.102716.
\bibitem[Ai et al.(2022)]{Ai} Ai J., Cai Y., Su Z., Zhang K.,Peng D., \& Chen Q. (2022). Predicting user-item links in recommender systems based on similarity-network resource allocation, \emph{Chaos, Solitons \& Fractals}, 158, 112032. https://doi.org/10.1016/j.chaos.2022.112032.
\bibitem[Chen et al.(2022)]{Chen} Cheng Y.L, Hsiao C.H., \& Wu C.C. (2022). An ensemble model for link prediction based on graph embedding, \emph{Decision Support Systems}, 157, 113753, https://doi.org/10.1016/j.dss.2022.113753.
\bibitem[Calderoni et al.(2020)]{Calderoni}Calderoni, F., Catanese, S., De Meo, P., Ficara, A.,\& Fiumara, G. (2020). Robust link prediction in criminal networks: A case study of the Sicilian Mafa
\emph{Expert Systems with Applications}, Vol. 161,113666. https://doi.org/10.1016/j.eswa.2020.113666.
\bibitem[Aref \& Wilson(2019)]{Aref2019} Aref, S. \& Wilson, M. C. Balance and frustration in signed networks. \emph{Journal of Complex Networks} 7, pp.163–189, https://doi.org/10.1093/comnet/cny015.
\bibitem[Aref et al.(2020)]{Aref2020} Aref, S., Mason, A. J. \& Wilson, M. C. A modeling and computational study of the frustration index in signed networks. \emph{Networks} 75, pp.95–110, https://doi.org/10.1002/net.21907.
\bibitem[Antal \& Krapivsky(2006)]{Antal} Antal, T., Krapivsky, P. L., \& Redner, S. (2006). Social Balance on Networks: The Dynamics of Friendship and Enmity, \emph{Physica D} 224, 130, https://doi.org/10.1016/j.physd.2006.09.028.
bibitem[Ribeiro et al.(2018)]{Ribeiro} Ribeiro, H. V., Alves, L. G. A., Martins, A. F., Lenzi, E. K. \& Perc, M. (2018). The dynamical structure of political corruption networks. \emph{Journal of Complex Networks} 6, pp.989–1003, https://doi.org/10.1093/comnet/cny002.
\bibitem[Arinik et al.(2019)]{Arinik} Arinik, N., Figueiredo, R. \& Labatut, V.(2019). Analysis of roll-calls in the European parliament by multiple partitioning of multiplex signed networks. \emph{Social Networks}, https://doi.org/10.1016/j.socnet.2019.02.001.
Reviewer 2 Report
The authors used deep learning machines to model the evolution of political signed networks created from the data produced by the Correlates of War project.
This work is interesting and easy-to-read. I have just few suggestions:
1) The introduction is very short and it does not provide sufficient background and references. In my opinion, since the authors wrote about the applications of link prediction algorithms in lines 19-20, they could add some details about this topic. For example, an interesting application is the one about criminal networks. In this regard, I can suggest the following reference: Calderoni, F., Catanese, S., De Meo, P., Ficara, A., Fiumara, G.: Robust link prediction in criminal networks: A case study of the Sicilian Mafia. Expert Systems with Applications 161, 113,666 (2020), https://doi.org/10.1016/j.eswa.2020.113666. There are also applications in other fields such as biology, medicine etc. which could be relevant. Also the paragraph on political networks is very short. Therefore, the authors could think about adding a Related work section in which they can produce a deeper discussion about the state-of-the-art on political temporal networks and the approaches used on them.
2) The Proposed Approach section could be split in two separate subsections. In the first subsection the authors could describe their approach, and in the second one they could make a deeper description of the dataset. In my opinion, a table could also help to understand the characteristics of the network such as the number of nodes, the number of edges and so on.
3) A minor spell check is required. For example, I noticed two typos in lines 56 (“approachs”) and 69 (“the the”).
Author Response
Point 1: Introduction too short: Please see lines 12-84 the introduction has been extended and the following references has been included:
\bibitem[Elhesha(2019)]{Elhesha}
Elhesha R., Sarkar A., Boucher J., \& Kahveci T. (2019) Identification of co-evolving temporal networks , \emph{BMC Genomics}, Vol. 20(Suppl 6):434, doi: 10.1186/s12864-019-5719-9. https://pubmed.ncbi.nlm.nih.gov/31189471/
\bibitem[Wang \& Gan(2021)]{Wang}
Wang W., Wuang L., \& Gan C. (2021). Epidemic Spreading Dynamics on Temporal Networks, Collection of papers 2021, https://www.hindawi.com/journals/complexity/si/813208/
\bibitem[Masuda(2017)]{Masuda}
Masuda N. , \& Holme P. (Eds.) (2017). \emph{Temporal Network Epidemiology (Theoretical Biology)} 1st Edition, Springer, https://link.springer.com/book/10.1007/978-981-10-5287-3.
\bibitem[Zhu et al.(2022)]{Zhu} Zhu, Z., Sun Y., Kuang Y., Yuan X., Gu H., Zhu J., \& Xing W. (2022). Contemporaneous symptom networks of multidimensional symptom experiences in cancer survivors: A network analysis. \emph{Cancer Medicine} Wiley, pp. 1-11. https://doi.org/10.1002/cam4.4904.
\bibitem[Jordan et al.(2020)]{Jordan} Jordan G., Winer S., \& Salem T. The current status of temporal network analysis for clinical science: Considerations as the paradigm shifts?. https:/onlinelibrary.wiley.com/doi/10.1002/jclp.22957.
\bibitem[Nasuha-Daud et al.(2020)]{Nasuha-Daud} Nasuha-Daud N., Hafizah-Ab-Hamid S., Saadoon M., Sahran F., \& Badrul-Anuar N.(2020). Applications of link prediction in social networks: A review, \emph{Journal of Network and Computer Applications}, Vol. 166, 2020, 102716. https://doi.org/10.1016/j.jnca.2020.102716.
\bibitem[Ai et al.(2022)]{Ai} Ai J., Cai Y., Su Z., Zhang K.,Peng D., \& Chen Q. (2022). Predicting user-item links in recommender systems based on similarity-network resource allocation, \emph{Chaos, Solitons \& Fractals}, 158, 112032. https://doi.org/10.1016/j.chaos.2022.112032.
\bibitem[Chen et al.(2022)]{Chen} Cheng Y.L, Hsiao C.H., \& Wu C.C. (2022). An ensemble model for link prediction based on graph embedding, \emph{Decision Support Systems}, 157, 113753, https://doi.org/10.1016/j.dss.2022.113753.
\bibitem[Calderoni et al.(2020)]{Calderoni}Calderoni, F., Catanese, S., De Meo, P., Ficara, A.,\& Fiumara, G. (2020). Robust link prediction in criminal networks: A case study of the Sicilian Mafa
\emph{Expert Systems with Applications}, Vol. 161,113666. https://doi.org/10.1016/j.eswa.2020.113666.
\bibitem[Aref \& Wilson(2019)]{Aref2019} Aref, S. \& Wilson, M. C. Balance and frustration in signed networks. \emph{Journal of Complex Networks} 7, pp.163–189, https://doi.org/10.1093/comnet/cny015.
\bibitem[Aref et al.(2020)]{Aref2020} Aref, S., Mason, A. J. \& Wilson, M. C. A modeling and computational study of the frustration index in signed networks. \emph{Networks} 75, pp.95–110, https://doi.org/10.1002/net.21907.
\bibitem[Antal \& Krapivsky(2006)]{Antal} Antal, T., Krapivsky, P. L., \& Redner, S. (2006). Social Balance on Networks: The Dynamics of Friendship and Enmity, \emph{Physica D} 224, 130, https://doi.org/10.1016/j.physd.2006.09.028.
bibitem[Ribeiro et al.(2018)]{Ribeiro} Ribeiro, H. V., Alves, L. G. A., Martins, A. F., Lenzi, E. K. \& Perc, M. (2018). The dynamical structure of political corruption networks. \emph{Journal of Complex Networks} 6, pp.989–1003, https://doi.org/10.1093/comnet/cny002.
\bibitem[Arinik et al.(2019)]{Arinik} Arinik, N., Figueiredo, R. \& Labatut, V.(2019). Analysis of roll-calls in the European parliament by multiple partitioning of multiplex signed networks. \emph{Social Networks}, https://doi.org/10.1016/j.socnet.2019.02.001.
Point 2: The proposed approach section splited in two separate sections. Please see lines 85-145.
Round 2
Reviewer 1 Report
The authors took into account all my comments and accordingly improved the text of the article, which is now of better quality and significantly improved. My recommendation is to accept this paper for publishing as it is.
Author Response
Two more references has been added to the introduction according to academic editor comments. Please see line 23
