Wind Power Forecasting Based on LSTM Improved by EMD-PCA-RF
Round 1
Reviewer 1 Report
Thank you for your effort.
I have some comments that have to be considered in the modified manuscript.
------------------------------
1) The paper deals with improving the accuracy of wind power forecasting. Please, show the significance of this issue clearly.
2) The (Abstract) contains 2 main results (numerical values) that cannot indicate the high performance of your work. Instead: how can I judge these values are good or suitable.
3) The (Conclusion) must include the main results of your work (numerical values).
4) The (Introduction) must be ended with a paragraph showing the paper organization.
5) I prefer to draw a block diagram (or a flow chart) to describe the procedure of the mathematical model. (NOT that of Fig. 3..... I mean only the mathematical or equations sequence).
6) Many equations in the model must have references. Ex: Eqs. (1) +(2)+(3).... is it Ref. 49?
7) In Table 5, what do you mean with (importance)?
8) The vertical axis in Fig. 7 needed to write its name (error).
9) I do NOT see any comparison with any published work. How can you judge your results are correct.
10) Table 10 needs adding a column (to the right) showing the percentage enhancement, when using different models.
11) Concerning references list, it is preferred to write all authors names. The abbreviation (et al.) can be used only within text, NOT in the references list. Please correct.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This study proposed a wind power forecasting model, namely EMD-PCA-RF-LSTM, incorporating Empirical Mode Decomposition (EMD), Principal Component Analysis (PCA), Random Forest (RF) and Long Short Term Memory (LSTM) neural networks. Overall, the manuscript is interesting, however, the following issues should be addressed.
- For Section 1, the authors should provide the comments of the cited papers after introducing each relevant work. What readers require is, by convinced literature review, to understand the clear thinking/consideration of why the proposed approach can reach more convincing results. This is the very contribution from the authors.
- The authors should clearly mention, point by point, the novelty of their work in section 1.
- In formula 1, please explain all the terms, including y and . Also, remove the miss typing from formula 1.
- Please properly write lines 142 to 144 in an academic style.
- In equation 10, replace the integration sign with a summation.
- It is highly recommended to provide the source of the data. In this way, other researchers will be able to reproduce your work or to compare their work with yours. The authors can upload the data set to some repository and share the link in the paper.
- Provide the time series plot of your response variable as well as explanatory variables. They will help the reader to see what specific features your time series data exhibit. For example, see (https://doi.org/10.1007/s12667-019-00356-w).
- The quality of the figures is too poor. The authors must improve the quality of the figures.
- The results listed in the tables should be explained in detail.
- For model selection, use adj-R square instead of R-square.
- It is not clear whether the results in Table 10 are for out-of-sample forecast or for in-sample?
- In addition, authors should also conduct some statistical tests to ensure the superiority of the proposed approach in Table 10, i.e., how could authors ensure that their results are superior to others? For example, see (Short-Term Electricity Demand Forecasting Using Components Estimation Technique).
- It is highly recommended to discuss other forecasting techniques used in energy markets in general to highlight the importance of forecasting techniques. For example, (Forecasting Natural Gas Spot Prices with Machine Learning) (Short-Term Electricity Prices Forecasting Using Functional Time Series Analysis)
- The conclusion must be shortened and precise.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
To: sustainability
From: Anonymous reviewer
Paper title: Wind power forecasting based on LSTM improved by EMD-PCA-RF
Subject: sustainability-1737420 peer-review report
Date: 26.05.2022
A brief summary
The authors’ main goal is to solve the problems of inaccurate feature identification and slow con- vergence in traditional wind power forecasting models, with full use of the environmental factor series that affect the wind power , mines the features of wind power and environmental factors over time.
Proposed EMD-PCA-RF-LSTM forecasting model to improve the forecasting accuracy of wind power, so as to provide certain technical support for the safe dispatch of power system and enhance the power system's ability for the consumption of wind power.
Dynamic time modeling of multivariate feature series is performed by LSTM (Long Short-Term Memory neural network) algorithm to forecast the wind power, and the model accuracy is verified by comparing with other forecasting models. Improving the accuracy of wind power forecasting can guarantee the stable dispatch and safe operation of the grid system
Wind power will keep developing rapidly and become an important part world’s power system construction in the future. Wind speed, temperature and other environmental factors, wind power has a strong randomness and fluctuation, large-scale wind power installation will bring greater challenges to the safe and stable operation of the power system and solving this problem is a significant step towards improving the efficiency of wind turbines.
The overall quality of the manuscript is good. The manuscript contains good-quality contents and results. The research methodology and conclusions presented in the paper can be used by designers and exploiters wind power systems.
The case study and discussion based on an actual China’s national wind energetic system are conducted. Numerous operational data were collected to accomplish this task.
The topic of the manuscript is also interest to the readers of sustainability.
However, the submission contains several technical and editorial drawbacks, which according to my humble opinion, require slightly revision. Some comments are given next.
Broad comments
Introduction - the problem analysis was correctly carried out based on rich and current literature sources. It should be noted, however, that of the 56 literature items, 49 are by Chinese authors.
A well conducted literature analysis allowing the reader to familiarise themselves with work and research methods carried out in other research centers.
Methods - methods used: LASSO algorithm, EMD algorithm, PCA-RF combined algorithm and LSTM algorithm are presented correctly with the mathematical apparatus used.
Data pre-processing and model design - properly designed the flowchart of the EMD-PCA-RF
LSTM based wind power forecasting model
Case study – the process: decomposition by: EMD, dimensionality reduction by PCA, feature extraction by RF, results and comparison and analysis of experimental results has been planned and carried out correctly
Conclusions - are correct and sufficient
Specific comments
It is appropriate to include in the paper a list of the abbreviations used
Line
374 Fig.4 is too small which makes it difficult to read the values obtained
395 Fig.5 - wrong description. Correct description should be: Factors of feature importance.
428 Fig.5 - wrong drawing description. This is not a comparison of models but of the results obtained using the different models
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors addressed my concerns and hence.
