Next Article in Journal
Economic Drivers of Voluntary Return among Conflict-Induced Internally Displaced Persons in Nigeria
Previous Article in Journal
Citizenship Ability, Homestead Utility, and Rural Homestead Transfer of “Amphibious” Farmers
 
 
Article
Peer-Review Record

Short-Term Prediction of PM2.5 Using LSTM Deep Learning Methods

by Endah Kristiani 1,2, Hao Lin 1, Jwu-Rong Lin 3, Yen-Hsun Chuang 3, Chin-Yin Huang 4 and Chao-Tung Yang 1,5,*
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Submission received: 21 December 2021 / Revised: 29 January 2022 / Accepted: 31 January 2022 / Published: 11 February 2022

Round 1

Reviewer 1 Report

Attached file

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

There are still some mistakes in language, as for example, in section 4.5.7

It says: The following table presents

must be: the following tables present

It says: the first 1 hour has more low RMSE values on average of 12

must be: the first 1 hour has lower RMSE with an average value 12

 

I will be willing to accept the explanation of extreme value processing in Fig 2, but I think it is still not convincing that Fig 14 shows LSTM as the best model. What is the statistical index under comparison in this graph? Is MAPE? If it is so, why it does not coincide with values of table 10?

I accept that the new figure 15 indicates that according to RMSE, LSTM is the best model.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Back to TopTop