NIDS via Data Analytics – [Hint]
- The input network transaction includes several
- These feature can be used to classify network transactions into normal or
- Just like we did in Spam Filtering to classify emails into spam or non-spam, you can build an IDS following the same approach.
- This means that your IDS model is just a classifier that can be training via a suitable
- As part of your research investigation, try to find suitable datasets for IDS
Examples of some Attacks (Intrusions)
Example of Transaction Features
- Depending on the dataset used, your problem can be modeled as binary-class or multiclass classification problem.
• Write (TYPED) a research proposal (1000+ words) on Intrusion Detection Systems (IDSs) using data analytics.
- Refer to for the good survey: Machine Learning and Deep Learning Methods for Intrusion Detection Systems: A Survey
• Your proposal should include the following sections:
- Title page
- Introduction [about 1 page]
- Literature review (each student in the group describes one paper) [0.5 page for each paper]
- Research Methods and Resources [about 75 page]
- Conclusion [0.5 page]
• Format required for your report
- Font type: Times New Roman
- Font size: 12pt
- Line spacing: 5-spaced
- Submission file format: PDF
- Microsoft Word Page layout: Letter (8.5’’ x 11’’)
- That’s approximately 5 pages (excluding figures, tables or references)
Choose the “Letter” page layout
- Title page [excluded from word count]
- Title of your proposal
- Group name
- Group members (Name, ID and Major)
- Try to interest the reader in the topic
- Define the problem you are trying to solve (what is intrusion detection?)
- What is network intrusion detection?
- Why it is important to solve this problem?
- What is data analytics and how it can solve this problem?
- Provide some historical/cultural context for the
- At the end of your introduction, include a tentative thesis to indicate to your reader that you are entering your project looking at your topic through a critical, analytic lens – this thesis should clearly state your intentions using a meta-discursive structure (g., “In this project, I will…” or “This research project will investigate …”).
- A good recipe for the introduction is context followed by (problem or complication) followed by proposed argument or research Each ingredient in this recipe should be 3-6 sentences long.
- Literature review (include at least 6 relevant papers)
- List how previous research targeted this problem
- Each student in the group should pick one paper and write:
- Main contributions
- Analyze it
- Critical review
- Research gap in the paper and how it can be improved
- Datasets used?
- If the group has 6 student, then the literature review should include 6 papers
- You can organize your literature review based on time period, topic, methods used,
- Create connections and transit from one idea to another
- DO NOT just list the contributions without connection
- At the end of the LR section, add a table to compare the papers you studied
• Research Methods and Resources
- Describe the methods you will be using to model and solve the intrusion detection problem via data analytics
- Describe the computer software to be used
- Describe the data analytics learning approach (supervised/ unsupervised …)
you will be using
- Describe the concept your model is trying to learn (classification or regression or clustering or association analysis or …)
- Describe the dataset(s) you will be using in building your data analytic learning model
- Describe the evaluation methods you will be using to evaluate your solution
• Conclusion [0.5 page]
- Address the lessons you learned from doing the literature review and learning about intrusion detection in computer systems.
- Address why should the reader want to read your article?
- Address any challenges you faced and how you are going to overcome them
in the next part of the project.
• Contributions breakdown
- List the names of the authors of each of the following contributions:
- Conceptualization: author 1, author 2, …
- Methodology: author 1, author 2, …
- Software: author 1, author 2, …
- Validation: author 1, author 2, …
- Formal analysis: author 1, author 2, …
- Investigation: author 1, author 2, …
- Resources: author 1, author 2, …
- Data curation: author 1, author 2, …
- Writing: author 1, author 2, …
- Editorial: author 1, author 2, …
- Visualization: author 1, author 2, …
- References [Include at least 9 references]
- Use a unified reference style
- Use the APA referencing and citation style