How Data Science Could help Cybersecurity Cases

Field that significantly transform because of Data Science

Photo by FlyD on Unsplash

You might have heard about it, but a big ransomware case is happening in Indonesia.

So, what happened is that the Indonesian National Data Center was attacked on 20 June 2024 by cyber terrorists using ransomware. In turn, around 200 government facility affected and disturbed significantly, including the Indonesian Immigration system.

The damage is substantial, not only in the monetary aspect but also in the reputation, as it turns out that the government hasn’t performed any data backup all these years, and the security is compromised because of a simple matter (password negligence).

This situation is certainly something that every company or country wants to avoid. While the Indonesian government has its negligence to blame for this situation, the cyberterrorist here is certainly at fault.

To prevent cyber attack cases like ransomware, data science can help in some use cases.

How does data science could help? Let’s explore the possibility.

Data Science and Cybersecurity

I remember playing the Watch Dogs game in my younger days, and my first thought was how easy it is for hackers to access our data. Like, the protagonist looks at the civilian face, and they have all our detailed data at once.

I know it’s just a game, but the possibility was there. Cyberterrorists could compromise our important data, just like what happened in Indonesia. That’s why the cybersecurity field was created to ensure that cases like ransomware don’t happen.

Data science seems far from cybersecurity, but data science has since become part of modern cybersecurity. Data science has become more important than ever by providing tools and techniques to defend against cyber threats.

For example, machine learning algorithms can be used for anomaly detection.

The anomaly detection algorithms can analyze patterns in network traffic, user behaviour, and system logs to identify anomalies that might indicate a security threat.

Image by Gabriel Pierobon

For example, you can see in the image above that an anomaly happens in the credit card transaction (high transaction amount). While customers may have unique one-time high transactions, it could indicate something else. That’s why credit card providers often call us if there is an unusual activity.

The anomaly detection algorithm would include, but not limited to:

1. Isolation forest

2. Local outlier factor

3. Robust covariance

4. One-class support vector machine (SVM)

5. One-class SVM with stochastic gradient descent (SGD)

We can extend anomaly detection for many cybersecurity activities, as mentioned above.

Speaking of anomalies, Network security is another area in cybersecurity where data science has helped significantly. Cybersecurity teams can better understand normal network behaviour and spot anomalies by applying time series analysis and clustering algorithms to network traffic data.

Graph analysis techniques can also map network connections and uncover unusual patterns. These methods allow for a more proactive approach to network defence, allowing organizations to prepare before being attacked by cyber threats.

Network Security with Graph Analysis

For example, the network graph can be used to visualize network activities and create rules allowing cybersecurity to act if any anomaly occurs.

The paper by Bonacina et al. (2020) also proposes combining time series, clustering, and network graphs as one methodology. It’s a great method that, in my opinion, can be used in the cybersecurity field.

Bonacina et al. (2020)

Overall, if we look at the future, data science will transform cybersecurity. Just look at the current Generative AI trend where many cyber scammers have implemented the model to create a convincing persona to scam innocent citizens.

Both defenders and attackers would increasingly leverage AI. In the middle of these would be data science, which keeps developing new technology that, unfortunately, can be used by both parties.

That’s all for now! I hope you enjoy my latest newsletter and spark something in you.

---

If you need any help or want me to write something about your interest, just comment or contact me on my social networks. Or even better, use the chat!

---

Articles to Read

Here are some of my latest articles you might miss this week.

NIS America

• 5 Tips for Getting Started with Deep Learning in Machine Learning Mastery.

• Optimizing Scikit-learn Models for Better Performance in Statology.

• Streamlining Your Machine Learning Workflow with Scikit-learn and Joblib in Statology.

• Using Scikit-learn’s Manifold Learning for Non-linear Dimensionality Reduction in Statology.

• Evaluating and Improving Model Robustness Using Scikit-learn in Statology.

---