Research on the Chaotic Characteristics and Noise Reduction Prediction of Information System Anomalies in Equipment Manufacturing Enterprises

Round 1

Reviewer 1 Report

This submission, as presented to me as a new reviewer, is a version of slight modification probably based on another round of revision. Authors seemed having added some more discussions to original manuscript.

However, as a general reader starting to read it for the first time, I still have trouble to find a clue of the overall structure of the paper. 

The section II seemed to be coming from nowhere. From section I, which was actually combining an introduction and a literature review section, but failing to specify the formulated problem in a formal style, section II started directly with all maths and a non-self-explanatory Figure 1 by throwing a huge amount of information like C-C and Wolf method, and suddenly a transition to focus on SVD. I am totally lost in finding a few things:

1. what is main problem? where is definition of the chaotics, noise, prediction, and what are objectives for applying RBF ANN?
2. and how such a AI problem was related to EME's IS? 
3. and how to follow on section III, where a case was thrown on readers faces without essential explanation of the main problem?

Consequently, it is even more difficult to understand what the data meant and what the experimental results would mean for other enterprises' further application of the proposed method. 

Authors might try to provide such information in section 4.2, but this set of newly added paragraphs could not answer my question above, in other words, how to build the problem formulation in a generic way so that the data, parameters, NN structure, or input/output/prediction metrics, can all be linked within one framework, so an enterprise with so-called chaotic abnormal IS system problems might benefit from such 'theoretical contribution' from this particular paper. There is a big gap between empirical value and jargon maths/ANN derivations. 

Authors need to be trained on how to present ideas, materials, and research results in a professional or mode academic way. 

Author Response

First of all, thank you for your valuable advice.

For your confusion and suggestions, we have made the following modifications:

(1). The method framework of the section 2, namely Figure 1, is described in detail (see Lines 149- 166), the whole 2.1-2.3 parts are modified (see Sections 2.1-2.3), the formulas are numbered (see equation (1)-(29)), and the symbols appearing are described (see Table A1 of Appendix A);

(2). The case companies are selected from representative ZZ enterprises, and the ERP abnormal data covers the whole process from single coverage stage to integration promotion stage, which we think can reflect the current situation of this kind of enterprises. The methods used in the research (whether it is chaos recognition, noise reduction or prediction) can be used as reference in their respective scopes. It is worth noting that the noise reduction method and prediction method need to adjust the parameters according to the actual data, so the method has certain universality.

Please see the attachment for specific reply to your questions 1, 2 and 3.

Author Response File: Author Response.pdf

Reviewer 2 Report

April 4, 2021

Review of the paper (sustainability-1148063-peer-review-v1)

Research on Chaotic Characteristics and Noise Reduction Prediction of Abnormality of Information System in Equipment Manufacturing Enterprises

by  Peng Niu, Yanming Sun and Zhuping Gong

 General Comments

In this paper, the Authors proposed abnormal index time series data for analysis of the integration process of two industries. In the first step, they have calculated the embedding dimension, time delay, average period and the maximum Lyapunov exponent of the time series. Then, as prediction tools radial basis function (RBF) neural network and local nonlinear method are applied. It turned out that one local noise reduction can dig hidden problems in the actual enterprises’ operation. The results obtained show that multiple iterations of local noise reduction can extract mainstream information of signals, avoid failure at isolated points and show a clear attractor structure. The paper presents interesting considerations and the results obtained are, in my opinion, promising.

Major Comments

I find the idea presented in the paper interesting and promising but I have the following comments and concerns: 

1. What is the size of the signal considered in this work?
2. In the section introduction the paragraph considering the role of noise in information transmission. Here are a few interesting and some important recent papers that could be useful for the preparation of such a brief analysis/comments:

- J. Zhang, D. Zhou, D. Cai, and A. Rangan, A coarse-grained framework for spiking neuronal networks: between homogeneity and synchrony, J. Comput. Neurosci. 37(1), 81–104 (2014).

- S.A. Oprisan, A. Sorinel, and C.V. Buhusi, Why noise is useful in functional and neural mechanisms of interval timing?”, BMC Neurosci. 14(84), 1–12 (2013).

- I. Kanitscheider, R. Coen-Cagli, and A. Pouget, “Origin of information limiting noise correlations,  Proc. Natl. Acad. Sci. U.S.A. 112(50), e6973–e6982 (2015).

- A. Pregowska, E. Kaplan, and J. Szczepanski, How Far can Neural Correlations Reduce Uncertainty? Comparison of Information Transmission Rates for Markov and Bernoulli Processes, Int. J. Neural Syst. 29(0), 1950003–1–13 (2019).

- A. Pregowska, Signal Fluctuations and the Information Transmission Rates in Binary Communication Channels, Entropy, 23, 92 (2021).

2. Section 2. Models and Methods should be widely described in terms of the presented research methodology framework (presented in Figure 1 without any comments!).
3. All equation should be numbered and these numbers should be referenced in the appropriate places in the main text. At this moment it is not easy to read/follow the paper.
4. Following d) Section 2.1-2.3 should be rewritten with presented more clearly. At the moment, it is unable to understand by the Readers.
5. Figure 3 is unclear in so a scale (time axis).
6. What conclusion follows from Table 2?
7. In the discussion the proposed methodology should be compared with already existed ones, for example, taking into account the accuracy. Maybe in the form of a Table, or comment in the main text.
8. It should be pointed out in the abstract what kind of accuracy Authors have in mind. It is only written as “good prediction accuracy”, it is not a precise term.
9. In the Abstract in the sentence “Finally, radial basis function (RBF) neural network and local nonlinear method are used to predict.” it is unclear to predict of what?
10. All (!) quantities and symbols used should be explained/defined (specifically, including that in formulas and the Tables!).
11. Figure 6. What is the meaning of parameter t? Is this a time ?! Why it is not written explicitly?
12. Figure 8. The axes are not described! The same for Figure 9 and Figure 10 and similarly horizontal axe in Figure 11.

In summary, the paper requires a significant improvement of its editorial form.

Minor Comments

1. The language should be improved definitely.
2. In line 277 there is unnecessary line breaking.

Final Comments

The idea presented and developed in the paper seems to be interesting and the results obtained are promising (i.e. they are of some value), but due to the above concerns, I do not recommend this paper for publication until the above comments/questions will be carefully addressed. At this moment I would recommend Major Revision.

Author Response

First of all, thank you for your very detailed and valuable comments.

We have revised your concerns and suggestions one by one.

Please see the attachment for specific reply.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Technically and logically it is much better

English polish is still needed

Author Response

Thank you again for your valuable comments.

The suggestion has been adopted, and this article has undergone English language editing by MDPI.

Please see the attachment for English Editing Certificate.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors address my comments and remarks, but one literature item is misspelt on the references list (it should be written: [25] Pregowska, Signal Fluctuations and the Information Transmission Rates in Binary Communication Channels, Entropy, 23, 92 (2021).) After the correction, I would recommend this paper for publication.

Author Response

Thank you again for your valuable comments.

The suggestion has been adopted, and the misspelled literature item in the reference list has been modified (see References 25.).