ICACT20220050 Question.1
Questioner: liumengup@126.com 2022-02-12 오후 9:08:23 |
ICACT20220050 Answer.1
Answer by Auhor zakaria.sawado@gmail.com 2022-02-12 오후 9:08:23
Chrome  Click!! |
| this paper investigated the impact that imbalanced datasets can have on the performance of models produced by classification algorithms. The analysis of the most suitable evaluation metrics for imbalanced dataset is proposed.
1.The train dataset and test dataset are both balanced or both imbalanced? that is, the situation that the train dataset is balanced and the test dataset is imbalanced is considered or not. this situation may occure frequently. |
Thank you for your interest in our research.
For your first question, we used two types of datasets (balanced data and imbalanced data).Then we formed training and test data for each type of data set.
Then we formed training and test data for each type of data set.
Your second question was whether we considered a balanced training set and an unbalanced test set, the answer is no. Thank you for this lead that we can explore.
|
ICACT20220050 Question.2
Questioner: tomayoon@ieee.org 2022-02-14 오후 7:39:17 |
ICACT20220050 Answer.2
Answer by Auhor zakaria.sawado@gmail.com 2022-02-14 오후 7:39:17
Chrome Click!! |
| First of all, thank you for the good research paper and presentation contents. I would like to know more about the research environment used in this study, and also in detail about the equipment & operating systems, development languages and software libraries you used. |
We conducted our experiment on an HP Z620 workstation with the following characteristics: Intel (R) Xeon (R) CPU E5- 2643 0 @ 3.30 GHz 3.30GHz (2 processors), RAM: 34 GB, on
which we installed Windows 10, 64 bit operating system. We used Anaconda Navigator which is a graphical user interface(GUI) which integrates several development tools such as
Jupiter, Spyder and allows the use of python 3.6 and Scikit-learn, a free Python library for machine learning |
ICACT20220050 Question.3
Questioner: jypark9740@g.skku.edu 2022-02-14 오후 7:48:46 |
ICACT20220050 Answer.3
Answer by Auhor zakaria.sawado@gmail.com 2022-02-14 오후 7:48:46
Chrome Click!! |
| I would like to know the detailed information about the data set used in the experiment. |
We used the CICMalDroid 2020 data from the Canadian Institute for Cyber Security. You can get more information at this website: https://www.unb.ca/cic/datasets/maldroid-2020.html |
ICACT20220050 Question.4
Questioner: nay.min@student.mahidol.ac.th 2022-02-15 오후 1:25:26 |
ICACT20220050 Answer.4
Answer by Auhor zakaria.sawado@gmail.com 2022-02-15 오후 1:25:26
Chrome Click!! |
| Do you think the current dataset being used is enough to evaluate your testing? And I would like to know if you have any plan to use other datasets for the evaluation in the future? |
Yes, we have used an acceptable dataset for our tests but this does not prevent us from evaluating in the future as new data is always important |
ICACT20220050 Question.5
Questioner: akmoon@etri.re.kr 2022-02-15 오후 1:38:53 |
ICACT20220050 Answer.5
Answer by Auhor zakaria.sawado@gmail.com 2022-02-15 오후 1:38:53
Chrome Click!! |
| What is the ratio of majority and minority of the unbalanced data used?
What data did you use for the experiment? |
We used a dataset composed of several five classes. As presented in figure 9 and for these classes we considered an unbalanced DataA and balanced DataB for our experiment.
We used the CICMalDroid 2020 data from the Canadian Institute for Cyber Security |
ICACT20220050 Question.6
Questioner: rna0415@g.skku.edu 2022-02-15 오후 4:19:34 |
ICACT20220050 Answer.6
Answer by Auhor zakaria.sawado@gmail.com 2022-02-15 오후 4:19:34
Chrome Click!! |
| Thank you for your work. I am also interested in an adversarial attack that perturbs classification. I just want to know about your work's limitations and following future goals (next research goal). Thank you in advance. |
Thank you for your interest
Our work does not deal with adversarial attacks but we have studied the impact of unbalanced datasets on the performance of machine learning models in the field of
android malware |
ICACT20220050 Question.7
Questioner: namacabale@gmail.com 2022-02-16 오후 1:38:01 |
ICACT20220050 Answer.7
Answer by Auhor zakaria.sawado@gmail.com 2022-02-16 오후 1:38:01
Chrome Click!! |
| Dear Author, you have a very good research. Is there an existing technology almost similar to your proposal that has been deployed in real world environment? What does it lack that your work specifically addressed? Thank you and stay safe.
|
To our knowledge there is little similar to our proposal but does not address the same issues as we do. We are working in to perfect our approach and deploy it in a real environment in the future. |
