An Efficient Cellular Automata-Based Classifier with Variance Decision Table

Round 1

Reviewer 1 Report

Paper-ID: applsci-2261510

Title: An Efficient Cellular Automata-Based Classifier with Variance Decision Table

Recommendation: Major Revision

This research work, suggest a classifier building using cellular automata and variance decision table. The contribution does not seem to be sufficient to be published in a journal. Validation of proposed work is insufficient. Thus the author is advised to revise the manuscript in this regard.

Major updates:

Ø  Accuracy is not the only measure to confirm your model. Thus take 3 to 4 more performance metrics such as sensitivity, specificity, gmean, jaccard to show your outcome.

Ø  Apply statistical method (such as Topsis) to prove your analysis.

Ø  Number of classification model considered for proposed analysis is also limited. There are many other machine learning and also deep learning model (such VGG16, VGG19,GNF, InceptionNet, DenseNet etc) Few out of these can be used to strengthen this paper.

Minor updates:

·         English language of the manuscript must be thoroughly check. Many typos, few statements are repeated etc.

·         Specify the parameters in Equn 14 clearly.

·         Why in Equn 18, it is [2-1] in the denominator.

·         Define Equn 20 then use it.

·         Use the following paper to strengthen your research findings.

o   Ban Y, Wang Y, Liu S, Yang B, Liu M, Yin L, Zheng W. 2D/3D Multimode Medical Image Alignment Based on Spatial Histograms. Applied Sciences. 2022; 12(16):8261. https://doi.org/10.3390/app12168261

o   Tripathy, Jogeswar, Rasmita Dash, Binod Kumar Pattanayak, Sambit Kumar Mishra, Tapas Kumar Mishra, and Deepak Puthal. 2022. "Combination of Reduction Detection Using TOPSIS for Gene Expression Data Analysis" Big Data and Cognitive Computing 6, no. 1: 24. https://doi.org/10.3390/bdcc6010024

o   Lu, S., Yang, B., Xiao, Y., Liu, S., Liu, M., Yin, L.,... Zheng, W. (2023). Iterative reconstruction of low-dose CT based on differential sparse. Biomedical Signal Processing and Control, 79, 104204. doi: https://doi.org/10.1016/j.bspc.2022.104204

Thus the overall review comment: Similar kind of research work is available in the literature. It is a kind of extension work. Thus significant contribution need to be made. Thus the paper is subjected for major revision. If all these suggested points will be taken care of successfully, it can be consider for publication

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

 

In this study, the authors proposed a novel algorithm called a cellular automata based classifier with a variance decision table (CAV) for different binary and multiclass classification tasks. The work is novel in that the authors proposes the use of butterfly optimization rather than genetic algorithm for the task of classification. However, the work can be improved in many ways. Some recommendations are as under:

 

1.      Equations must be revisited.

2.      It is hard to understand the article due to statements like “using 10 K-fold cross-validations”.

3.      Lemma 4 and 5 on page 9 are the same statements.

4.      Cite the following articles appropriately:

·         https://doi.org/10.1007/s10278-019-00265-5

·         https://doi.org/10.3390/s22124609

·         https://doi.org/10.3390/su142214695

·         https://doi.org/10.3390/math9233101

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

REVIEW REPORT

 

Journal name: Applied Sciences (ISSN 2076-3417)

Manuscript ID: applsci-2261510

Title: “An Efficient Cellular Automata-Based Classifier with Variance Decision Table”

 

Comment for the authors:

 

In this paper authors proposed a cellular automata and variance decision table (CAV) based classifier algorithm to solve the high dimensionality, complexity, ambiguous, and inconsistent problem of datasets. The manuscript can be considered interesting for Applied Science readers; however, it needs significant improvements before being considered for publication.

 

Please, find my detailed consideration below.

 

·         The linguistic of the paper can be further improved. There are some typos. Some comma and spaces between words are missed. The language used throughout the manuscript needs to be improved

·         The motivation of this paper could be clarified. It is important to show the reason for doing this research.

·         Improve the introduction part by making it more concise.

·         The algorithm of the proposed method is well defined. However, the discussion part need to be improved.includ.

·         What is your fitness function? what is the outcome of the proposed algorithm?

·         The motivation of the proposed approach needs further clarification.

·         The main contribution and originality should be explained in more detail. It is a good way to show the advantage of the proposed method.

·         In the section 5 (Results and discussion) the authors should give the parameter Settings of the proposed algorithm and the comparison algorithm.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Paper-ID: applsci-2261510

Title: An Efficient Cellular Automata-Based Classifier with Variance Decision Table

Recommendation: Minor Revision

·         Few of the review comments are successfully updated and few are not.  Try to consider all comments.

·         For performance comparison, one deep learning model is implemented. It is insufficient.  Use 1 or 2 model to compare the outcome.

·         Do not write “Deep learning”, rather specify name of the model.

Author Response

Response to Reviewer 1 Comments - Round 2

Point 1: Few of the review comments are successfully updated and a few are not. Try to consider all comments.

Response 1: For a previous response to reviewer 1's comments, we have comprehensively answered and revised the manuscript comprehensively point by point from point 1 to point 3 of the major updates, and from point 4 to point 7 of the minor updates.

 

Point 2: For performance comparison, one deep learning model is implemented.  It is insufficient. Use 1 or 2 models to compare the outcome.

Response 2: Since all datasets are tabular. We provide two Dense Neural Networks which is one kind of deep learning, called DNN-1 and DNN-2, respectively, for comparison. DNN-1 and DNN-2 are different in activation function, number of hidden layers, and number of neurons in each hidden layer as shown in the paper (Table 2).

 

Point 3: Do not write "Deep Learning", rather specify the name of the model.

Response 3: We write DNN-1 and DNN-2 to replace "Deep Learning".

Author Response File: Author Response.docx