A Conceptual Design Specification Based on User Aesthetic Information Analysis and Product Functional Reasoning
Round 1
Reviewer 1 Report
This paper deals with the subject of user experience of industrial products. It focuses on improving the design by integrating functional and aesthetic design specifications based on the user experience of target users. The authors use neural networks to create a model for design specifications.
The work is interesting and well presented.
My comments are:
In lines 280 – 281 it is not clear if the classification of appearance specifications comes from literature, or it is authors’ decision. Please clarify this.
Lines 410-425: this is the major issue of this work. Evaluation surveys have always a level of subjectivity but the condition of participants having experience with at least one of the cameras is impossible to be satisfied. How did authors ensure this? How where the participants selected and which was the procedure that made valid the 300 responses?
Figure 9: After the optimization of the candidates, the value “display” increased from 0 to 1. Could the authors explain it?
A general comment about sections 4.5 and 5 is that they do not significantly differ, and I believe they should be rewritten.
Please revise the paper for syntax or other errors.
e.g. lines 84 – 85: looks like there is something missing (that ?)
lines 522 – 523: needs rephrasing
lines 409 – 414: the paragraph is repeated in lines 419 – 425.
Figure 7: Usability parameter “portability” is written “porpobility”
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 2 Report
I think the authors have submitted an interesting manuscript which is generally well written. The contribution is clearly stated and the description of the proposed method is clear except at one point.
I have the following comments and concerns:
i) The authors should create subsections within the "Introduction" section, such as "Contributions" and "Structure of the paper" to increase the readability of the manuscript.
ii) The structure of the applied neural network remains unclear from Figure 6 and from the description, as well. How many layers were in the NN? How many neurons were in each layer? What kind of optimizer was used during the training? Since training a neural network involves a lot of experiments, the publication of the training curves would be also nice.
iii) The authors applied genetic algorithms to improve the design specifications of Candidate 1. Why did the authors choose a genetic algorithm? Could simulated annealing, pattern search, or particle swarm optimization be also a good choice? Did you consider other optimization methods? I think the answer to these questions could further increase the scientific value of the manuscript.
Author Response
Please see the attachment.
Author Response File: Author Response.docx