Density Peaks Clustering Algorithm Based on a Divergence Distance and Tissue—Like P System

Round 1

Reviewer 1 Report

In this study, the authors proposed TP-DSDPC. Thus, in order to precisely estimate local density and relative distance of each point, a novel distance measure is introduced in TP-DSDPC. The clustering centres are chosen by score value automatically. The whole algorithmic process is carried out by a tissue-like P system. According to authors, TP-DSDPC outperforms other comparison algorithms when using a variety of synthetic and real-world datasets, according to three evaluation metrics.

My observations are listed below:

1.     The section on related work is deceptive. In fact, it omits to show how the authors' work compares to other works. You should revise it and provide a comparison table in this section.

2.     The language needs to be improved.

3.     Please refrain from using words like "last year," "essay," etc.

4.     Please make the figures 1, 2, and 3's design better. They appear very basic.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper proposes an improved density peaks clustering algorithm based on Divergence distance and tissue-like P system, named TP-DSDPC. It designs a divergence distance measure to estimate each point's local density and relative distance accurately. The distance relieves the impact of similarity measurement and the chain reaction. The algorithm first calculates the local density and updates the relative distance by divergence distance. Then, it automatically selects the clustering centers by the score value from high to low. Finally, it clusters the remaining points by the relative distance.

The experiment was conducted on nine datasets and compared the performance of the proposed algorithm with K-means, DBSCAN, DPC, KNN-DPC, DGDPC, and TP-DSDPC using several external evaluation metrics such as ACC, NMI, and ARI.

In general, the paper is well-written and well-organized. The topic is interesting and suitable for the scope of the journal. The experimental results seem to be reasonable. In support of this paper, the authors should consider the following points to improve the paper's quality further.

- In the Introduction, the authors should give some info on several other clustering methods. I have used k-means, DBSCAN, and HDBSCAN- an improved version of DBSCAN that allows varying density clusters instead of using a global epsilon distance as in DBSCAN. I observed that k-means, DBSCAN, and HDBSCAN could perform very well in the clustering task. In some cases, DBSCAN and HDBSCAN were even better than k-means since they can remove noises. I also used the HDBSCAN python version. It is an efficient algorithm in terms of runtime and can work well with high-dimensional data, arbitrary data shapes and identify outliers in the data. Thus, the authors should also summarize/compare the strength and weaknesses of HDBSCAN [https://hdbscan.readthedocs.io/en/latest/comparing_clustering_algorithms.html]. In addition, discuss possible methods that can perform clustering for an unknown number of clusters and provide high interpretability, such as hierarchical clustering; the authors can refer [https://doi.org/10.1007/978-981-15-1209-4_1] in the discussion.  From that, highlight the main advantages of the proposed method over other methods.

- In section 2, put a table of main notations used in the paper.
- In section 3,  it is better to give a figure showing the proposed framework's workflow. The workflow should include all steps from the input to the output.
- In section 3, theoretically discuss the complexity of TP-DSDPC.
- In section 4, put the standard deviation std for results shown in Tables 2, 3, 4, 6, 7, 8.
- Proofread the paper to fix all typos. There are many places in the paper the authors put no white space before "(" or "[". For instance, "(KNN-DPC) [20]" instead of "(KNN-DPC)[20]"; "Table 4 severally (the bolded" instead of "Table 4 severally(the bolded", etc.
I donot want to see this problem in the revised version!

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I'm satisfied. Thanks

Author Response

Thank you very much.

Reviewer 2 Report

Figures 3 and 4 are not in good shape; the authors should revise them to increase the resolutions and bigger and more copiable text.

Author Response

Please see the attachment.

Author Response File: Author Response.docx