Macro SOStream: An Evolving Algorithm to Self Organizing Density-Based Clustering with Micro and Macroclusters
Round 1
Reviewer 1 Report
In the manuscript, the authors have proposed a clustering algorithm named Macro SOStream, based on SOStream. The algorithm is based on self-organizing density for data stream clustering. The main features of algorithm are the creation of macro clusters with arbitrary shapes from micro clusters,
A merging mechanism for macro clusters to improve their performance and robustness to noisy data. The SOStream is based on Self-Organizing Map. Moreover, authors have pre-selected DenStream to compare with the proposed algorithm. This comparison is performed by analyzing the performance of the algorithms using the Adjusted Rand Index metric and the execution time of the data analysis.
It is concluded that the macro cluster merging functionality proved to be very relevant for the Macro SOStream algorithm’s good performance. Furthermore, execution time of Macro SOStream is better than DenStream.
The manuscript is well written and research is well organized.
Author Response
Dear reviewer,
We have read all of your comments on this work. We are very thankful and do appreciate all of them.
Thank you.
Reviewer 2 Report
This study presented a new approach Macro SOStream algorithm. The Authors bring a great contribution to the related field. However, this study can be perfect if the research limitation/ constraint can be included in the manuscript..
Author Response
Dear reviewer,
We have read all of your comments and suggestions on this work. We are very thankful and do appreciate all of them. In this letter, we answered of the addressed question, and we point to you the modifications performed in the text.
1 - However, this study can be perfect if the research limitation/ constraint can be included in the manuscript:
We thank you for your comment. Therefore, we added the limitations on page 19, lines 483 to 488. In the text, we commented on the use of only synthetic data for validation, and the execution time of the proposed algorithm, which is longer than the others.
Thank you.
