A Deep Ensemble Learning Method for Effort-Aware Just-In-Time Defect Prediction
Round 1
Reviewer 1 Report
No comments
Author Response
Dear Reviewer,
We appreciate you for spending time to review our paper and providing the valuable comments. It is your valuable and insightful comments that led to possible improvements in the current version. The author has carefully considered the comments and tried his best efforts to address every one of them. The author welcomes further constructive comments if any. We provided the point-by-point response as follows:
Sincerely,
Saleh Albahli
Reviewer 1:
Minor spell check is advised
The whole paper is revised with respect to style, grammar and language.
Reviewer 2 Report
After the revision, the paper looks more complete.
I still have some minor concerns on the presentation:
all figures should be 300DPI, otherwise, it looks blurred your deep neural network is called a multilayer perceptron (mention it) Figures 3 & 4 the two gigantic figures look ugly. Please put them side by side Table 2 the width is too large
The paper can be published when these minor changes are made.
Author Response
Dear Reviewer,
We appreciate you for spending time to review our paper and providing the valuable comments. It is your valuable and insightful comments that led to possible improvements in the current version. The author has carefully considered the comments and tried his best efforts to address every one of them. The author welcomes further constructive comments if any. We provided the point-by-point response as follows:
Sincerely,
Saleh Albahli
Reviewer 2:
Moderate English changes has been needed
We have revised the whole manuscript in order to correct the English and grammar related mistakes.
The neural network is call multilayer perceptron and you should mention it
We have replaced the name of DNN with multilayer perceptron where necessary.
Fig 3&4 looks ugly, paste them side by side.
All figures have been well placed as advised by changing them from .png to .tiff DPI300
Table 2 is large in width.
The table is revived to make it appropriate with respect to size and width.
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
This paper applies machine learning techniques to JIT software defect prediction. This is an application paper, meaning that no new method is proposed. Rather, the author applies an ensemble of three different classifiers to the target classification problem.
I recommend rejection based on the lack of a clear presentation of the problem and the poor presentation quality.
At the beginning of Section 3, the author has to clearly introduce the classification problem. How the input features are built, and what is target classes.
As ADAM optimizer is mentioned, it should be cited as well.
"Deep Neural Network" is too broad. The author should mention clearly what is the architecture of the neural network.
The presentation quality is poor, with low resolution figures, equations without explanations, and badly formatted table.
Overall I don't feel this work is ready to be published.
Reviewer 2 Report
The author proposes a prediction model based on a set of three individual classifiers applied to a Just-in-Time prediction.
The quality of the article is too poor, it is easy to follow and is well structured by in general is really poorly described:
* In the introduction it is not described the hypothesis, objectives, the context of applications. Only describe very general Just-in-Time methods and their problems.
* The literature review is short and it does not include the weakness of this area and how the new methods are going to go further over the state of the art.
* The methodology is poorly described, with a short enumeration of the methods, but without a detailed description of each method.
* The evaluation is produced over a public dataset.
So, in general, the article is not ready for its publication.
