All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
I am pleased to inform you that your work has now been accepted for publication in PeerJ Computer Science.
Thank you for submitting your work to this journal.
[# PeerJ Staff Note - this decision was reviewed and approved by Sebastian Ventura, a PeerJ Computer Science Section Editor covering this Section #]
Authors have incorporated the changes suggested. The paper may be accepted for the possible publication.
Authors have incorporated the changes suggested. The paper may be accepted for the possible publication.
NA
NA
The topic is interesting and the research presented is of adequate quality. I had concerns that were addressed by the authors.
The authors have given due attention to previous concerns.
The authors have given due attention to previous concerns.
After the prediction of each component, these are added to obtain the final forecast of the original signal. The current study is interesting. In general, the main conclusions are supported by the figures and text.
I have received reviews of your manuscript from three scholars who are experts on the cited topic. They find the topic very interesting; however, several concerns must be addressed regarding experimental results and design, and comparisons with currently existing methods. These issues require a major revision. Please refer to the reviewers’ comments listed at the end of this letter, and you will see that they are advising that you revise your manuscript. If you are prepared to undertake the work required, I would be pleased to reconsider my decision. Please submit a list of changes or a rebuttal against each point that is being raised when you submit your revised manuscript.
Thank you for considering PeerJ Computer Science for the publication of your research. We appreciate your submitting your manuscript to this journal.
This paper proposes a deep learning solution to the financial market forecasting.
There are few observations/suggestions to the authors.
1. Introduction section must end with the contributions of the proposed work.
2. Brief explanation on the dataset used is necessary.
3. As I understand this is a regression problem. Authors are suggested to include the error measures (Mean Absolute Error, Root Mean Square Error etc).
4. Good analysis is performed in the experimental section. The performance evaluations metrics must be applied if possible (namely accuracy and so on).
5. If possible, a comparative analysis of the proposed work vs existing work may be included.
This paper proposes a deep learning solution to the financial market forecasting. The article may be accepted for the possible publication after incorporating the suggestions mentioned.
The authors propose a method to forecast financial transactions by applying empirical mode decomposition to financial time series. Each intrinsic function is forecast with a back-propagation neural network. Then, all the predicted individual components are combined to obtain the overall predicted signal.
The use of EMD, instead of other signal decompositions, makes sense since financial time series are non-stationary and EMD can extract the main trend of these signals and suppress noise; however, EMD suffers from a major drawback, mixing mode. A better choice (suggestion for future work) would have been the use of other EMD versions that cope with such problem: CEEMDAN (complete ensemble empirical mode decomposition with added noise).
The number of days to forecast a future value depends on a sliding window of different sizes (10, 20, 30, 40, 50, 60, 70, 80, 90, 100). Are individual IMF forecasts made on actual input data within the sliding window or on previous predicted values within the window? Are IMF forecasts estimated in a single-day or multiple-day way? This part requires more explanations.
There is not information regarding the architecture of the neural network (number of layers and number of neurons per layer). How did the architecture was determined? These parts require more explanations.
No comment.
Questions arise regarding the architecture of the neural network, how hyper-parameters were chosen, and the process to forecast individual IMFs.
The paper studies the financial stock market forecasting problem and proposes the EMD based method.
There are some major problems as follows.
1. In the introduction, the long paragraphs describes the background of a wide range of financial market forecasting research, and the research is not focused enough. It should be more concise, focusing on the research work of stock forecasting.
2. The motivation for the proposed method is unclear. Why use the EMD based method?
3. The description of the method is unclear. In the method part, from the given Figure 1, the EMD method should be introduced first, then BNNN, and then PSO. The current structure should be revised.
The experimental results are not convincing enough. The experimental results are not compared with the current new method, and it is difficult to evaluate the effectiveness of the method.
No explicit findings are reported.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.