Framework of Meta-Heuristic Variable Length Searching for Feature Selection in High-Dimensional Data
Round 1
Reviewer 1 Report
The manuscript is interesting and falls into the scope of the journal. However, it lacks some shortcoming needed to be addressed before further consideration.
The abstract is short, especially when presenting the contribution of the paper, its importance, and a short conclusion.
The methodology is short, and it does not provide details information. For instance, which platform did you use to implement the model?
The results section is rather short, without providing an error analysis on the performance of the model. Additionally, the text is weak as the authors did not spend enough time to enrich the manuscript. also, the results need to be better discussed.
The manuscript lacks a discussion section where the application of the model in the field, its importance for the future research, and the future directions are discussed.
the conclusion section is also weak. The significance of the work should be discussed.
Author Response
The manuscript is interesting and falls into the scope of the journal. However, it lacks some shortcomings needed to be addressed before further consideration.
- The abstract is short, especially when presenting the contribution of the paper, its importance, and a short conclusion.
- The abstract has been updated The literature contains various meta-heuristic algorithms with variable length searching. All of them enables searching in high dimensional problems. However, an uncertainty in the performance of them exists. In order to fill this gap, this article proposes a novel framework for comparing various variants of variable length searching meta-heuristic algorithms in the application of feature selection. For this purpose, we we implemented four types of variable-length meta-heuristic searching algorithms, namely, VLBHO-Fitness, VLBHO-Position, variable length particle swarm optimization (VLPSO) and genetic variable length (GAVL) and we compare them in terms of classification metrics. The evaluation has shown the overall superiority of VLBHO over the other algorithms in terms of accomplishing lower fitness values when optimizing mathematical functions of the variable length type.
- The methodology is short, and it does not provide details information. For instance, which platform did you use to implement the model?
- We added this to the experimental evaluation section, For evaluation, we used MATLAB 2020b.
- The results section is rather short, without providing an error analysis on the performance of the model. Additionally, the text is weak as the authors did not spend enough time to enrich the manuscript.
- The manuscript will is to English-proofed.
- also, the results need to be better discussed.
- More elaboration of the results is provided
- The manuscript lacks a discussion section where the application of the model in the field, its importance for the future research, and the future directions are discussed.
- Applications and future directions are added
- the conclusion section is also weak.
- The conclusion is extended
- The significance of the work should be discussed.
- This research opens the door to adopt and adapt VLBHO for applying it in various areas of optimization research when the decision space does not fix length such as wireless sensor network deployment, data gathering, and variable length feature selection.
Reviewer 2 Report
This paper presents Framework of Meta-Heuristic Variable Length Searching for Feature Selection in High Dimensional Data. The research question is important and interesting. However, there are still some issues that should be addressed:
1. In the first part of the paper, the author writes “We are interested in two methods, namely, variable length genetic and variable length particle swarm optimization for feature selection”. However, in the conclusion the author writes “The study has considered three methods for the evaluation, namely, genetic variable length, variable length particle swarm optimization, and variable length black hole optimization with its two modes: position and fitness”. The main content of the paper is inconsistent.
2. In the fourth part of the paper, which function is used for comparing the algorithms.
3. What are the meanings of these terms: rosenbrock, rastrigin, sphere and griewank.
4. There is no reference [30].
5. What is the meaning of Table 1? There are no relevant explanations in the context.
6. There are also some mistakes such as Table II, Equations (9,10).
7. it is advisable to provide the full form before providing the abbreviated form.
Author Response
This paper presents Framework of Meta-Heuristic Variable Length Searching for Feature Selection in High Dimensional Data. The research question is important and interesting. However, there are still some issues that should be addressed:
- In the first part of the paper, the author writes “We are interested in two methods, namely, variable length genetic and variable length particle swarm optimization for feature selection”. However, in the conclusion the author writes “The study has considered three methods for the evaluation, namely, genetic variable length, variable length particle swarm optimization, and variable length black hole optimization with its two modes: position and fitness”. The main content of the paper is inconsistent.
- Thanks for your comments. we changed the first sentence to “We are interested in three methods for the evaluation, namely, genetic variable length, variable length particle swarm optimization, and variable length black hole optimization with its two modes: position and fitness”
- In the fourth part of the paper, which function is used for comparing the algorithms.
- The evaluation is conducted on four functions, namely, rosenbrock, Rastrigin, rastrigin , and sphere.
- What are the meanings of these terms: rosenbrock, rastrigin, sphere and griewank.
- They represent benchmarking mathematical functions used to evaluate optimization algorithms (https://doi.org/10.1007/s00500-020-04758-2). This reference is added.
- There is no reference [30].
- Reference added
- What is the meaning of Table 1? There are no relevant explanations in the context.
- It provides an overview of meta-heuristic searching algorithms which supports variable length searching. We added this to the introduction
- There are also some mistakes such as Table II, Equations (9,10).
- They have been checked
- it is advisable to provide the full form before providing the abbreviated form.
- All abbreviations are detailed in the article
Round 2
Reviewer 1 Report
The manuscript can be further considered for publication.
Reviewer 2 Report
Accept in present form