Mobile Phone Recommender System Using Information Retrieval Technology by Integrating Fuzzy OWA and Gray Relational Analysis
Round 1
Reviewer 1 Report
The authors propose a mobile recommender system based on the integration of fuzzy OWA with gray relational analysis.
The paper is quite interesting, but its novelty is limited and the related presentation and technical qualities have to be strongly improved for a possible publication.
First of all, the paper needs a linguistic revision.
In addition, the Introduction section should be enriched with a motivating example for the work that better highlights importance of the proposed recommender systems and with a part that describes the effective novelty and innovation of the proposed framework w.r.t. the state of the art.
The related work misses several relevant work on on hybrid, social and multimedia recommendations approaches that can address in different ways the proposed problem (see for example works on social recommendation by Moscato, Picariello, Colace and Albanese).
Because you are introducing a recommendation framework, its overview (integrating the entire recommendation process and the main system components) with the adopted technologies should be introduced.
The section discussing effectiveness performances should be enriched experimenting the proposed method on different type of collaborative filtering approaches (also w.r.t. the sparsity of ratings’ matrix) and by a part describing efficiency of the approach with respect to dataset size.
Author Response
The authors propose a mobile recommender system based on the integration of fuzzy OWA with gray relational analysis.
The paper is quite interesting, but its novelty is limited and the related presentation and technical qualities have to be strongly improved for a possible publication.
Response:
We have revised the manuscript according to the reviewers' and academic editor's comments.
First of all, the paper needs a linguistic revision.
Response:
We will use MDPI provides an English editing service checking grammar, spelling, punctuation and some improvement of style.
Manuscript (ID: english-6859 and Choice Language Service) has undergone English language editing by MDPI and Choice Language Service (as attached file).
In addition, the Introduction section should be enriched with a motivating example for the work that better highlights importance of the proposed recommender systems and with a part that describes the effective novelty and innovation of the proposed framework w.r.t. the state of the art.
Response:
(1)In the page 2, section 2, paragraph 4 added the RS literatures [22] and [23].
(2) In the page 4, above section 3:
This research integrated fuzzy OWA and gray relational analysis to design an intelligent RS structure to measure the factors affecting the target, such as quantification (cost, profit, etc.) and semantic level (satisfaction). The research has also considered the fuzzy theory generally in different decision-making situations [16, 20, 21, 22, 23]. The normalized fuzzy weight method [17] and modified Delphi method [2, 18] only consider the degree of semantics and cannot specifically quantify specific consideration factors, including cost, profit, etc., so the integration method of this research is both specific and practical.
The related work misses several relevant work on on hybrid, social and multimedia recommendations approaches that can address in different ways the proposed problem (see for example works on social recommendation by Moscato, Picariello, Colace and Albanese).
Response:
(1)In the page 2, section 2, paragraph 4 added the RS literatures Colace et al. [24].
Because you are introducing a recommendation framework, its overview (integrating the entire recommendation process and the main system components) with the adopted technologies should be introduced.
Response:
We are introducing a recommendation framework, its overview as below:
(1)In the page 1, section 1, paragraph 2.
(2) In the page 4, above section 3:
This research integrated fuzzy OWA and gray relational analysis to design an intelligent RS structure to measure the factors affecting the target, such as quantification (cost, profit, etc.) and semantic level (satisfaction). The research has also considered the fuzzy theory generally in different decision-making situations [16, 20, 21, 22, 23]. The normalized fuzzy weight method [17] and modified Delphi method [2, 18] only consider the degree of semantics and cannot specifically quantify specific consideration factors, including cost, profit, etc., so the integration method of this research is both specific and practical.
(3)In the pages 4~7, Section 3.
The section discussing effectiveness performances should be enriched experimenting the proposed method on different type of collaborative filtering approaches (also w.r.t. the sparsity of ratings’ matrix) and by a part describing efficiency of the approach with respect to dataset size.
Response:
(1)In the page 11, above table 5:
In this study, 30 experimenters were selected to carry out the experiment to measure the index of RS efficiency. The scholars had to have purchased mobile phones online on at least two occasions and be in the field of decision-making.
(2)In the page 12, below table 5:
The objective of this research was limited. Due to limitations in time and manpower, this study used convenient sampling to select the sample that best fit the purpose of this research (based on which experimenters had shopped online to purchase mobile phones). The main contents of the questionnaire included information quality, system quality, and service quality for overall satisfaction analysis. The questionnaire was conducted in an academic seminar in 2018. A total of 131 questionnaires were sent, and 100 valid questionnaires were collected.
(3)In the page 12, above section 5:
According to the calculation shown in Equation (7), the ratio of recall and precision is high, therefore the ratio of F1 can be increased. Due to the large number of functions of mobile phones and the different expectations of them, the reported precision and recall are particularly low, however the integration of fuzzy OWA and gray relational analysis shows that the precision ratio and F1 are better than other methods, therefore it can be used as a reference for the mobile phone product recommendation prototype system.
Reviewer 2 Report
The authors study a recommender system by integrating the weights of fuzzy order weighted averaging (OWA) and gray relational analysis. The authors’ proposed research framework calculates the recommended indices of the three weight calculation methods to be respectively 20.5%, 14.36%, and 16.43% after 30 experimenters’ examination. The authors’ proposed research framework is of practical interest to improve the performance of the recommender systems that are utilized nowadays especially for the mobile phones. The proposed scheme seems to have low complexity (needs to be clarified) and on the other hand improved achieved performance. However, the authors should address the following comments to improve the presentation and concreteness of their manuscript. 1) Initially, the provided literature review is misleading as the authors do not discuss the most well-known methods that support the recommender systems nowadays and contribute to the personalization and social recommendation mechanism (e.g., "Personalized multimedia content retrieval through relevance feedback techniques for enhanced user experience." In Telecommunications (ConTEL), 2015 13th International Conference on, pp. 1-8. IEEE, 2015, "Person identity label propagation in stereo videos", IEEE Trans. Multimedia, vol. 16, no. 5, pp. 1358-1368, 2014, etc.). The section of literature review (Section 2) needs to be substantially rewritten and discuss the existing methods that are used to the existing recommender systems. The references list needs to be updated with more relevant research works to the authors’ research. 2) Following my previous comment, there exists the content-based recommendation, the collaborative-based recommendation and the hybrid recommendation approach to offer to the end-users the most relevant content (e.g., A holistic approach for personalization, relevance feedback & recommendation in enriched multimedia content. Multimedia Tools and Applications, 77(1), pp.283-326, 2018, Content-based, collaborative recommendation. Commun ACM 40(3): 66–72, 1997, Recommender systems survey. knowledge-based systems. Elsevier 46:109–132, 2013, etc.). The authors should refer to the main differences among those methods and how their proposed recommender system is differentiated compared to them, or adopts parts of the already existing approaches to further improve them. 3) The authors should provide some additional comparative numerical results to other relevant research works from the recent literature in order to show the benefit of adopting their proposed framework. 4) The complexity analysis and implementation cost of the presented framework should be provided and discussed in the numerical results. 5) The presented numerical results show mainly the pure performance of the proposed framework, however without a comparative study to other approaches, the reader is not able to understand the benefits of the proposed framework. 6) Also, the whole manuscript should be carefully checked for typos, grammar and syntax errors, as there are many of them. 7) One additional question that is raised based on the authors’ provided analysis, is how the number of the attributes that the proposed framework uses is being determined? If the authors set a small number of attributes, how the corresponding recommended results are influenced? 8) The authors consider within their analysis a recommended threshold. It is not clear if this threshold is related to the end-users personal characteristics. If it is, the authors should clarify how it is related? If it is not, the authors should explicitly state that. The manuscript needs a minor revision before acceptance.
Author Response
The authors study a recommender system by integrating the weights of fuzzy order weighted
averaging (OWA) and gray relational analysis. The authors’ proposed research framework calculates the recommended indices of the three weight calculation methods to be respectively 20.5%, 14.36%, and 16.43% after 30 experimenters’ examination. The authors’ proposed research framework is of practical interest to improve the performance of the recommender systems that are utilized nowadays especially for the mobile phones. The proposed scheme seems to have low complexity (needs to be clarified) and on the other hand improved achieved performance. However, the authors should address the following comments to improve the presentation and concreteness of their manuscript.
Response :
(1) In the page 11, above table 5:
In this research, 30 experimenters were selected to carry out the experiment to measure the index of RS efficiency. The scholars had to have purchased mobile phones online on at least two occasions and be in the field of decision-making.
(2) Mobile phones have many functions.
(3) According to (1) and (2) this research provides complex solutions.
1) Initially, the provided literature review is misleading as the authors do not discuss the most wellknown methods that support the recommender systems nowadays and contribute to the
personalization and social recommendation mechanism (e.g., "Personalized multimedia content
retrieval through relevance feedback techniques for enhanced user experience." In Telecommunications (ConTEL), 2015 13th International Conference on, pp. 1-8. IEEE, 2015, "Person identity label propagation in stereo videos", IEEE Trans. Multimedia, vol. 16, no. 5, pp. 1358-1368, 2014, etc.). The section of literature review (Section 2) needs to be substantially rewritten and discuss the existing methods that are used to the existing recommender systems. The references list needs to be updated with more relevant research works to the authors’ research.
Response 1:
(1)In the page 3, section 2, paragraph 4 added the RS literatures Pouli et al. [25] and Zoidi et al. [26].
2) Following my previous comment, there exists the content-based recommendation, the collaborative-based recommendation and the hybrid recommendation approach to offer to the end-users the most relevant content (e.g., A holistic approach for personalization, relevance feedback & recommendation in enriched multimedia content. Multimedia Tools and Applications, 77(1), pp.283-326, 2018, Content-based, collaborative recommendation. Commun ACM 40(3): 66–72, 1997, Recommender systems survey. knowledgebased systems. Elsevier 46:109–132, 2013, etc.). The authors should refer to the main differences among those methods and how their proposed recommender system is differentiated compared to them, or adopts parts of the already existing approaches to further improve them.
Response 2:
In the page 3, section 2, paragraph 4 added the RS literatures [27], [28] and [29].
3) The authors should provide some additional comparative numerical results to other relevant research works from the recent literature in order to show the benefit of adopting their proposed framework.
Response 3:
In the page 11, below the Eq. (7):
This research used the Normalized fuzzy weight [30] and Modified Delphi method [31] to compare the Integrating fuzzy OWA and gray relational analysis, as shown in Table 5.
4) The complexity analysis and implementation cost of the presented framework should be provided and discussed in the numerical results.
Response 4:
(1)In the page 11, above table 5:
In this study, 30 experimenters were selected to carry out the experiment to measure the index of RS efficiency. The scholars had to have purchased mobile phones online on at least two occasions and be in the field of decision-making.
(2)In the page 12, below table 5:
The objective of this research was limited. Due to limitations in time and manpower, this study used convenient sampling to select the sample that best fit the purpose of this research (based on which experimenters had shopped online to purchase mobile phones). The main contents of the questionnaire included information quality, system quality, and service quality for overall satisfaction analysis. The questionnaire was conducted in an academic seminar in 2018. A total of 131 questionnaires were sent, and 100 valid questionnaires were collected.
(3)In the page 14, above section 5:
According to the calculation shown in Equation (7), the ratio of recall and precision is high, therefore the ratio of F1 can be increased. Due to the large number of functions of mobile phones and the different expectations of them, the reported precision and recall are particularly low, however the integration of fuzzy OWA and gray relational analysis shows that the precision ratio and F1 are better than other methods, therefore it can be used as a reference for the mobile phone product recommendation prototype system.
5) The presented numerical results show mainly the pure performance of the proposed framework, however without a comparative study to other approaches, the reader is not able to understand the benefits of the proposed framework.
Response 5:
In the page 11, below the Eq. (7):
This research used Normalized fuzzy weight[30] and Modified Delphi method [31]compared the Integrating fuzzy OWA and gray relational analysis as shown in the table 5.
6) Also, the whole manuscript should be carefully checked for typos, grammar and syntax errors, as there are many of them.
Response 6:
We will use MDPI provides an English editing service checking grammar, spelling, punctuation and some improvement of style.
Manuscript (ID: english-6859 and Choice Language Service) has undergone English language editing by MDPI and Choice Language Service (as attached file).
7) One additional question that is raised based on the authors’ provided analysis, is how the
number of the attributes that the proposed framework uses is being determined? If the authors set a small number of attributes, how the corresponding recommended results are influenced?
Response 7:
(1) In the pages 8~11, shown in the section 4.1.
(2)In the pages 11~12, shown in the section 4.2.
8) The authors consider within their analysis a recommended threshold. It is not clear if this threshold is related to the end-users personal characteristics. If it is, the authors should clarify how it is related? If it is not, the authors should explicitly state that. The manuscript needs a minor revision before acceptance.
Response 8:
In the page 8, section 4, step 7 is explain as below:
The mobile phone product materials whose conformity ratios (include three methods) are greater than the threshold values set by the users independently are selected and ranked, and the results are recommended to the users. Because the end-users personal characteristics are different of research. This research had higher requirements for product functions, therefore the threshold values were set to 85%, as shown in Table 5.
Round 2
Reviewer 1 Report
The authors addressed the rquested revisions. The paper is in a good shape for a possible publication.
Response:
The authors addressed the rquested revisions. The paper is in a good shape for a possible publication.
