Editor Comments

Associate Editor
Comments to the Author:
The authors are advised to carefully address the several comments from the first reviewer and the two concerns from the second reviewer while going through a minor revision round for this paper.  Please summarize the changes made in response to the comments from both reviewers in a separate document.

********************

Reviewer Comments

Please note that some reviewers may have included additional comments in a separate file. If a review contains the note “see the attached file” under Section III A – Public Comments, you will need to log on to ScholarOne Manuscripts to view the file. After logging in, select the Author Center, click on the “Manuscripts with Decisions” queue and then clicking on the “view decision letter” link for this manuscript. You must scroll down to the very bottom of the letter to see the file(s), if any.  This will open the file that the reviewer(s) or the associate editor included for you along with their review.

Reviewer: 1

Recommendation: Author Should Prepare A Minor Revision

Comments:
The paper proposes a framework for profiling demand in cellular networks. Similarity between traffic patterns is used to classify network usage based on K-means clustering. Multiple stopping rules are aggregated for deciding the optimal number of clusters. The authors performed two detailed case studies on large-scale datasets, with careful analysis on the cause of outliers. The paper shows enough differences to the conference version. It is well written, well organized and easy to understand. I have the following minor concerns. 

1. In my opinion, cognitive networking is the potential application area for the proposed demand profiling network (e.g., demand-based resource allocation). In fact, the two case studies are based on data from cellular networks. The Wi-Fi network is part of mobile network, but it is not included in the analysis, which I understand is due to a lack of data. So the title may be changed to "Mobile Demand Profiling of Mobile (or Cellular, which is more appropriate) Networks", and corresponding changes should be made to abstract and introduction. To say mobile demand profiling for cognitive network is overreaching as no specific features of cognitive radio network is considered in modelling. 

2. I am a little confused by the calculation of the top-k index. It is said that it is based on the evaluation of the k best clusterings. For example, In Fig. 4, Top-10 index is used. But when the actual number of clusters is fewer than 10 (this is indeed the case when it is decided that 2 or 3 clusters is the best), how do we select 10 best clusters? I guess that if the number of clusters is, say, 3, we select the whole 3 clusters instead of 10. Is this true? If so, please explain clearly in the paper. 

3. The legend in Fig. 5 and Fig.9 is not in order. I see no reason to name the clusters as "C1 C2 C0 C3" instead of "C0 C1 C2 C3". 

4. Minor grammar mistakes of using "As for" (page 10, left column line 52 and page 11 right column line 6). The authors can use "Just like" or "The same as". 

5. Page 11, left column, line 32, Class C2 actually includes night to early morning (18:00 to 3:00 +1 day) instead of only (18:00 to 0:00)?

Additional Questions:
1. Which category describes this manuscript?: Practice / Application / Case Study / Experience Report

2. How relevant is this manuscript to the readers of this periodical? Please explain under Public Comments below. : Very Relevant

1.  Please explain how this manuscript advances this field of research and/or contributes something new to the literature. : This paper propose a framework to profile typical network-wide demand in the cellular network, with detailed analysis on unusual situations. Different from the existing demand prediction models, the proposed model is un-parametric. The proposed clustering methods are applied to two case studies on large-scale real datasets, with careful analysis on outliers.

2. Is the manuscript technically sound? Please explain under Public Comments below. : Yes

1. Are the title, abstract, and keywords appropriate? Please explain under Public Comments below. : No

2. Does the manuscript contain sufficient and appropriate references? Please explain under Public Comments below. : References are sufficient and appropriate

3. Does the introduction state the objectives of the manuscript in terms that encourage the reader to read on? Please explain under Public Comments below. : Yes

4. How would you rate the organization of the manuscript? Is it focused? Is the length appropriate for the topic? Please explain under Public Comments below. : Satisfactory

5. Please rate the readability of this manuscript. Please explain your rating under Public Comments below. : Easy to read

6. Should the supplemental material be included? (Click on the Supplementary Files icon to view files): Yes, as part of the digital library for this submission if accepted

7. If yes to 6, should it be accepted: As is

8. If this manuscript is an extended version of a conference publication, does it offer substantive novel contributions beyond those of the previously published work(s)- i.e. expansion of key ideas, examples, elaborations etc.  *New results are not required*: Yes

Please rate the manuscript. Please explain under Public Comments below. : Excellent


Reviewer: 2

Recommendation: Author Should Prepare A Minor Revision

Comments:
In this work, the authors propose a framework which builds mobile demand profiles based on their analysis on mobile operator data. Snapshots are extracted from mobile traffic data. The authors then definite and compute two similarity measures - traffic volume similarity and traffic distribution similarity to compare different traffic patterns. A hierarchical algorithm(UPGMA) is adopted to do the clustering, and they explained stopping rules in detail so as to precisely separate clusters. They evaluate it on two real-world datasets, extracting useful traffic patterns and state the potential connection between mobile demand profiles and special events. In general, the idea is novel and interesting and the work is complete.
 
This paper is the first to design mobile traffic analytics for the large-scale data collected within a cellular network. It successfully shows that specific patterns of mobile traffic profiles can be effectively studied. The authors also do a survey of several different stopping rules of the clustering algorithm, which makes the work more complete and significant. The method proposed in this paper is meaningful for it may contribute to proper deployment of cognitive network resources in future. Meanwhile, it is also the first to separate distinguishable mobile demand profiles in a target region. The results of two case studies are impressive.
 
To further improve this paper, the authors are suggested to consider the following comments.
1.      The number of categories is too small compared to the vast variation of demand profiles in time and space. How to introduce a more fine-grained discrete resolution? Will the improvement be time-consuming?
2.      While the connection between outliers and mobile demand profiles is convincing, there is no surprise for the readers. More interesting discussions are expected.

Additional Questions:
1. Which category describes this manuscript?: Research/Technology

2. How relevant is this manuscript to the readers of this periodical? Please explain under Public Comments below. : Relevant

1.  Please explain how this manuscript advances this field of research and/or contributes something new to the literature. : In this work, the authors propose a framework which builds mobile demand profiles based on their analysis on mobile operator data. Snapshots are extracted from mobile traffic data. The authors then definite and compute two similarity measures - traffic volume similarity and traffic distribution similarity to compare different traffic patterns. A hierarchical algorithm(UPGMA) is adopted to do the clustering, and they explained stopping rules in detail so as to precisely separate clusters. They evaluate it on two real-world datasets, extracting useful traffic patterns and state the potential connection between mobile demand profiles and special events. In general, the idea is novel and interesting and the work is complete.

2. Is the manuscript technically sound? Please explain under Public Comments below. : Yes

1. Are the title, abstract, and keywords appropriate? Please explain under Public Comments below. : Yes

2. Does the manuscript contain sufficient and appropriate references? Please explain under Public Comments below. : References are sufficient and appropriate

3. Does the introduction state the objectives of the manuscript in terms that encourage the reader to read on? Please explain under Public Comments below. : Yes

4. How would you rate the organization of the manuscript? Is it focused? Is the length appropriate for the topic? Please explain under Public Comments below. : Satisfactory

5. Please rate the readability of this manuscript. Please explain your rating under Public Comments below. : Easy to read

6. Should the supplemental material be included? (Click on the Supplementary Files icon to view files): Does not apply, no supplementary files included

7. If yes to 6, should it be accepted: As is

8. If this manuscript is an extended version of a conference publication, does it offer substantive novel contributions beyond those of the previously published work(s)- i.e. expansion of key ideas, examples, elaborations etc.  *New results are not required*: Yes

Please rate the manuscript. Please explain under Public Comments below. : Excellent