Skip Content
You are currently on the new version of our website. Access the old version .
MDPIMDPI
Search all articles
by
  • Gloria Bordogna*,
  • Paola Carrara and
  • Simone Lella
  • et al.

Reviewer 1: Adnan Yazici Reviewer 2: Anonymous Reviewer 3: Anonymous

Round 1

Reviewer 1 Report

The paper proposes a novel functionality for a Geographic Information Retrieval (GIR) system that simplifies the process of searching for and visiting different types of geo-resources in a geographic area. This is achieved by retrieving and ranking several routes visiting a number of relevant georeferenced resources as a result of a single query, named flexible trip-planning query. The proposed retrieval model takes into account personal user preferences to rank the most convenient routes.

The paper also presents a graph-based algorithm that utilizes prioritized aggregation to optimize the routes' identification and ranking. The proposal is applied in the proof-of-concept of a Smart cOmmunity-based Geographic infoRmation rEtrievAl SysTem (SO-GREAT), which collects and manages open data of regional authority describing categories of authoritative territorial resources and services, such as schools, hospitals, etc., and Volunteer Geo24 graphic Services (VGSs) created by citizens to offer services in their neighborhood.

Overall, the paper presents an interesting approach to simplifying the process of searching for and visiting geo-resources in a geographic area. The proposed retrieval model and graph-based algorithm have the potential to significantly improve the user experience in this regard. The proof-of-concept application of the proposed functionality in SO-GREAT demonstrates the feasibility of the approach and provides a solid foundation for further research in this area. However, the paper could benefit from a more detailed explanation of the proposed retrieval model and graph-based algorithm, as well as a more comprehensive evaluation of the proposed approach.

Author Response

We thank reviewer 1 for her having appreciated the pros of our paper and we tried to meet the suggestions; in particular, we have introduced  more explanations and refinements of the graph-based algorithm by recalling its main objectives, and discussing a running example.

With respect to a more detailed explanation of the proposed retrieval model, the meaning of this request is not clear, as a great part of the description of the implementation is devoted to explaining and exemplifying how information are searched and retrieved.

With respect to a more comprehensive evaluation of the proposed approach, we are going to perform a specific test to this purpose and a specific publication will discuss the results; we have now specified this in the paper conclusions. 

Reviewer 2 Report

An approach, named flexible trip-planning queries, is proposed. It aims at retrieving and ranking several routes visiting a number of relevant georeferenced resources. To this end a graph-based algorithm is defined, exploiting prioritized aggregation to optimize the routes’ identification and ranking. The proposal is applied in the proof-of-concept of a Smart cOmmunity-based Geographic infoRmation rEtrievAl SysTem (SO-GREAT) designed to strengthen local communities.

The theoretical foundations of the approach are sound and semantically clear. The topic addressed is very interesting and the approach is promising.

However, the paper could be improved w.r.t. many aspects:

- In section Related Work, the content provided seems complete and exhaustive. 

On the other hand, it would be of a great interest for the readers and researchers to add a comparative table of the works reviewed w.r.t some criteria conveniently chosen (or at least categorize such works in families or classes). 

- Figure 1 is illisible and in this state, it does not bring any insight. Idem for the other figures.

- The notations used is not easy to understand. Illustrative examples are needed from section 2.4. At least provide a running example with some details showing some calculus. For instance, one does not how the parameters p and r_rank in Table 2 are obtained, idem for the factor RSV in Tables 4 and 5.

- In section 3.2.3, examples are also needed to better illustrate the data model and the routes function.

- Finally, is the approach costly? Please explain.

 

 

Author Response

We thank reviewer 2 for the comments on our proposal.

With respect to the punctual comments:

  • As suggested we added in the related work section a paragraph with a main categorization of the approaches;

“To recap the main characteristics of the reviewed trip planning approaches we consider these two main categories:

  • approaches for best routing to relevant resources, which require to identify the relevant territorial resources belonging to categories of interest declared in the query [9];
  • best itinerary planning approaches, which require to specify the starting and ending point of the trip and a cost function such as minimum distance, minimum time, and/or qualitative aspects of the roads specified by query keywords [23-28]. “

 

  • We have reduced the figures to only those meaningful and enlarged the remaining for a better reading.
  • We simplified notation and added explanation of the calculus in the captions of the running examples

 

 

  • We added the running example first introduced in subsection in subsection 2.3, then evaluated in subsection 3.1.
  •  

We discussed the complexity of the algorithm after introducing the graph-based algorithm

Reviewer 3 Report

1.      Instead of the traditional road selection algorithm based on minimum travel distance or time, this paper intends to propose a method for ranking the suggested route based on users’ preference with ordered visited places. The authors develop a web-based system, SO-GREAT, to demonstrate the designed architecture.

2.      While many factors are considered, the discussion about how SO-GREAT (Section 3.2) prepares and handles the data being used is too general and difficult to establish necessary links with the proposed methods and examples in section 3.1. For example, while uses with different background (preferences) are considered, the analysis of ranked suggested routes should be presented along with real examples, but the discussion is rather limited. I do suggest the discussion in section 3 should be reorganized to present the data being tested, the processing of data being considered (e.g., short texts from VGSs or terms like flexible metric conditions), the assessment of ranking for different background of users, etc.

3.      The discussion of 9-intersection needs formal definitions about the interior, boundary and exterior of the objects. How to justify the numbers being used to define the various scenarios of in_neighborhood. A figure showing the different scenarios would be helpful.

4.      Although the goal is to find route with good suggestion of visiting order based on users’ preference, the location aspect (i.e., coordinates) shall still play a role in the ranked result, can this aspect also be discussed?

5.      Can the data from VGSs provide “sufficient information” for making suggestions? VGI typically have to consider uncertainty and incomplete information, how it is processed and used and how it impacts the outcomes need to be discussed.

6.      For the arguments “ the routes are ranked”, can this be justified with the examples discussed in section 3?  How to explain or prove the ranked result is the one with the best suggested order. This also relates to the chosen fuzzy set membership function and should be further explained and assessed.

7.      The suggestion outcomes does require a correct interpretation of users’ preference, how to deal with situations where users cannot clearly specify their request?

8.      The quality of figures need to improve, some of them are hard to read.

9.      The style of reference also need to check, some of them are not consistent. Some of them are missing. (LINE1147-1149)

Author Response

We thank reviewer 3 for the comments and suggestions. Hereafter we answer punctually.

2- as far as subsection 3.2 describing how data are collected and prepared and linked to the data model we introduced connections with the other sections. We have explained the links with the graph based algorithm and where it is used and when describing the routes ranking definition in subsection 2.4.1 we introduced links to the implementation subsection 3.2.3.

3- We provided definition of interior exterior and boundary of spatial objects and explained the heuristic behind the definition of the in_neighbourhood operator by adding a new figure.

4- the role of geographic coordinates is played when evaluating the spatial conditions close  and in-neighbourhood between spatial objects and thus it plays a role in determining the RSV of the trip. We have recalled and explicitly explained this.

5-       This depends on how rich, accurate and complete is the description of the VGS provided by the volunteer. A textual field is provided that the volunteer can fill with a long textual description. Besides, the comments on the VGS in the tweets of other users can provide useful information for evaluating the trustfulness of the VGS. When discussing the role of the tweets filtering this aspect is discussed.

6-       Yes the routes are ranked based on the prioritised aggregation operator results. This is stressed in the final version.

7-       The visiting order is assumed as the order of the types of interest listed in the query. Flexibility is provided to identify the types by means of thesauri and by the definition of the soft constraints on the spatial conditions allowing to cope with the drop out problem caused  by crisp conditions using thresholds. This has been added in the revised version.

8-       We reduced the figures to only those meaningful and added a new figure for the 9-intersection model.

9-       We revised the format of references.