The paper performs feature selection using Antlion optimization (ALO) with s- and v-shaped transfer functions. The results indicate that the v-shaped transfer function is suitable with ALO, and it generated better results than GSA and PSO. Overall, the paper is written well but it also maintains significant opportunities for improvement.

Following are my observations about this paper: 1- Title suggests that the purpose of research is competition between v- and s-shaped transfer functions when applied with ALO on feature selection problem. 2- The abstract clearly explains the problem being addressed in this research. However, it does not match with the title which highlights that S- vs. V-shaped transfer functions to be applied on Ant Lion Optimization (ALO). Since, the title infers some sort of competition between the two transfer functions, there is no outcome which may refer to any of the function being suitable and the other unsuitable for ALO in feature selection problems. The research has proposed 6 versions of ALO on what basis, is also untouched in this section. Also, the abstract must mention some statistics like the percentage of overall feature selection and overall accuracy achieved in this study. 3- Some of the references are older like [4] and [9] which can be replaced with latest research. The style of writing references is also inconsistent and in come cases wrong like [1], [2], [8], [19], [22], etc. 4- The introduction section introduces the research topic properly, along with similar other works already reported in literature. The authors support the importance of their research in this section, and the logical progression of of current knowledge is also well presented. Later, the purpose of this work is also well explained, which is applying six variants of S- and V-shaped transfer functions on ALO to convert into its binary variant for solving feature selection problem. 5- There is minor punctuation error in the end of Introduction section, it ends with comma instead of a period. 6- In Section 3, the Algorithm 1 has some format issues. For statement should be ended with End For. The statements after inner For statement are not in bullet form as the ones given in inner For. 7- In this study, K-NN is used as classifier which is not an efficient algorithm and maintains certain drawbacks. SVM could also be used in this study to ensure validity of the selected subset of features. 8- The datasets used in this study are with small number of features. Feature selection research meaningful only when it is applied on datasets with feature-size more than 500. 9- The discussion about Table 5 should be placed after discussion on Table 4. 10- Table 4 should contain figures in same number digits after decimal, to increase readability. For example, 6, 5.6, 7.33 are three different formats which can be replaced with one format as 6.00, 5.60, 7.33. Please refer to Table 3 for consistancy in format. 11- The results section must end with discussion on solidly explaining the reasons why ALO-V3 performs better than other V-shaped transfer functions and also in comparison with S-shaped. How this particular transfer function produces better search results must be backed. 12- A graphical representation of the results is lacking in this section. The average accuracy, number of selected features, and computational times could be more easily illustrated in charts. 13- This conclusion Section lacks limitations of the study. The authors should explain limitations of using S- and V-shaped transfer functions.