Traffic Flow Prediction for Smart Traffic Lights Using Machine Learning Algorithms
Round 1
Reviewer 1 Report
In this paper, the use of machine-learning (ML) and deep learning (DL) models are proposed for the traffic flow prediction at an intersection
to dynamically adjust the optimal times of the states in the traffic lights. A public dataset is used for the training and testing of the proposed models.
- Use the abbreviations like machine learning (ML) and deep learning (DL). This is more like a standard usage form.
- The authors use only 56 says data with a min frequency. In my view, this data set is quite small. It should at least be one year to cover the whole year as patterns usually can vary in different seasons.
- Some paragraphs need formatting according to the journal recommend guidelines e.g check paragraph 2 on page 1. Also spacing between paragraphs.
- Figure 1 needs to be redrawn as the axis are not labeled and it seems like a snapshot. Put the whole image here. Also, in figure 3, the y-axis is not labeled.
- There is no need to put the links in the paper e.g. the libraries or software. just put a reference number here and details of the web pages in the reference list. e.g 30, 31, 32,
- The first paragraph in the results needs to revise, there is no need to mention so many references here. It is unnecessary.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 2 Report
From the paper: "models are proposed for the traffic flow prediction at an intersectionto dynamically adjust the optimal times of the states in the traffic lights "
the paper shows data prediction for time series. But there is nothing here about the optimal switching times for traffic lights. And traffic lights will obviously change the traffic itself.
There is also no analysis of the robustness of the proposed models.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper is presenting the using of machine learning to predict the traffic flow for smart traffic lights.
The authors are using a Dataset from Huawei Munich Research Center. They split the data from the dataset into two sets, one for training and one for the test.
The results are very well explained, maybe you can improve the explication of the figures. You can help the reader to understand the figure more easily.
I propose also, to improve the section of the proposed used scenario because is not very clearly explained.
The references are new and on the topic.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The paper is in much better shape now. Mostly the changes have been incorporated in the new version. The paper may be accepted for publication now.
Reviewer 2 Report
All my comments have been taken into account