Journal of Artificial Intelligence Research and Applications

By Scientific Research Center, London 534

AI in Fraud Detection: Leveraging Real-Time Machine Learning for

Financial Security

Ravi Teja Potla

Department Of Information Technology, Slalom Consulting, USA

Abstract

In an era where digital transactions dominate the global economy, fraud detection has become
a cornerstone of financial security. Traditional fraud detection systems, which rely heavily on
rule-based methodologies, are increasingly being outpaced by the sophisticated techniques
employed by modern fraudsters. These legacy systems struggle with adapting to the fast-
evolving landscape of digital fraud, often producing a high number of false positives and
suffering from delayed detection. As financial transactions increase in both volume and
complexity, the demand for more agile, accurate, and real-time fraud detection systems is
paramount.

This paper delves into the use of real-time machine learning models in revolutionizing fraud
detection. Unlike static, rule-based systems, real-time machine learning models can
continuously learn from vast datasets, identifying suspicious activities as they happen. These
models not only improve detection speed but also significantly reduce false positives,
ensuring that legitimate customer transactions remain uninterrupted. By leveraging
supervised learning models like Random Forests and Gradient Boosting Machines for
classification tasks, as well as unsupervised learning techniques like Autoencoders for
anomaly detection, machine learning can process large transaction datasets with
unprecedented efficiency.

We explore the challenges involved in implementing real-time machine learning models for
fraud detection, such as ensuring scalability, reducing system latency, and maintaining data
privacy. Furthermore, this paper addresses the ethical considerations surrounding AI-driven
fraud detection, particularly in terms of transparency and accountability. The integration
of Explainable AI (XAI) techniques into fraud detection models provides financial

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 535

institutions with the ability to understand and explain the decisions made by machine
learning algorithms, thus improving trust and compliance with regulatory frameworks.

The future of fraud detection lies in hybrid AI systems that combine various machine learning
models, reinforced with real-time data processing capabilities and enhanced by blockchain
technology for greater security and transparency. This paper concludes by discussing the
potential of these hybrid models to redefine fraud detection, ensuring financial institutions
are better equipped to combat increasingly sophisticated fraudulent activities.

Keywords:

Artificial Intelligence, Fraud Detection, Machine Learning, Financial Security, Real-Time

Analytics, Anomaly Detection, Predictive Analytics, Transaction Monitoring, Cybersecurity,
Fraud Prevention, Risk Management, Deep Learning, Behavioral Analysis, Financial Crime,
Money Laundering Detection, Credit Card Fraud, Security Automation, Pattern Recognition,
Data Privacy, Financial Institutions.

2. Introduction

The rise of digital transactions, mobile banking, and e-commerce has led to a significant
increase in fraud-related activities. Financial institutions are under pressure to detect
fraudulent transactions in real time to protect customers and prevent financial losses.
Traditional fraud detection systems, which rely on rule-based approaches, are increasingly
inadequate for handling complex and adaptive fraud patterns. These systems are often too
rigid to adapt to new types of fraud, resulting in delayed detection and high rates of false
positives, which can disrupt legitimate customer transactions.

To address these challenges, financial institutions are turning to real-time machine learning
models. These models are capable of processing large volumes of transaction data
instantaneously, identifying suspicious patterns, and predicting fraud as it occurs. Unlike
traditional systems, machine learning models can learn from historical fraud data and
continuously adapt to emerging fraud techniques. The ability to detect fraud in real time,
while minimizing false positives, is revolutionizing financial security.

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 536

This paper explores how real-time machine learning models are transforming fraud detection
systems. We will examine various machine learning approaches used in fraud detection,
including supervised learning, unsupervised learning, and reinforcement learning. We will
also discuss the technical challenges involved in deploying real-time models at scale, such as
maintaining low system latency and ensuring data privacy. Furthermore, this paper outlines
future directions in fraud detection, including the integration of Explainable AI
(XAI) and blockchain technology to enhance security and transparency.

3. Current Challenges in Fraud Detection Systems

Despite advancements in AI, current fraud detection systems face several inherent challenges:

3.1 Latency in Decision-Making

Traditional fraud detection systems struggle to process large amounts of transaction data in
real time. Transactions are often delayed while systems analyze the data, resulting in a poor
customer experience. Moreover, by the time a fraudulent transaction is flagged, the damage
may already be done. Real-time fraud detection aims to address this issue by processing data
as transactions occur, providing instant analysis and action.

3.2 High False Positive Rates

Rule-based systems often generate an overwhelming number of false positives. While these
systems can catch many instances of fraud, they frequently flag legitimate transactions as
suspicious, disrupting customer service and leading to frustration. Machine learning models,
particularly those using unsupervised techniques, can help reduce false positives by learning
more complex patterns in transaction data.

3.3 Adaptive Fraud Techniques

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 537

Fraudsters are constantly evolving their techniques, making it difficult for static systems to
keep up. As new types of fraud emerge, financial institutions must constantly update their
rule sets, which can be time-consuming and ineffective. Real-time machine learning models
offer a dynamic solution, as they can adapt to new fraud patterns without the need for manual
intervention.

4. Leveraging Real-Time Machine Learning for Fraud Detection

4.1 Real-Time Data Processing

Real-time machine learning enables financial institutions to process large datasets instantly,
identifying fraudulent transactions as they happen. This section will explore how data streams
from transactions are fed into real-time machine learning models, which analyze the data and
flag suspicious activities. The ability to handle vast amounts of data in milliseconds allows
financial institutions to stay ahead of fraudsters.

Figure 1: Real-Time Machine Learning Workflow for Fraud Detection

This workflow diagram illustrates the real-time process of fraud detection. Transaction data
is processed by machine learning models such as Random Forests and Autoencoders, which
analyze the data and produce a fraud detection output. The integration of these models into
financial systems enables near-instantaneous detection of suspicious activities, improving
both detection accuracy and response time.

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 538

4.2 Machine Learning Models in Fraud Detection

Several machine learning models are commonly used in real-time fraud detection:

• Supervised Learning: Algorithms such as Logistic Regression, Random Forests,

and Gradient Boosting Machines are often used for fraud classification tasks. These
models are trained on historical data to distinguish between legitimate and fraudulent
transactions.
• Unsupervised Learning: Algorithms like Autoencoders and Isolation Forests are
employed when labeled fraud data is scarce. These models detect outliers or unusual
behavior that could indicate fraud.
• Reinforcement Learning: In some cases, reinforcement learning models are used to
improve fraud detection over time. These models are designed to learn from previous
actions and rewards, continuously optimizing the detection strategy.

4.3. Examples of Real-Time Machine Learning Techniques

Real-time machine learning models are designed to process data streams in milliseconds,
enabling institutions to detect fraud as transactions occur. Some of the most effective
techniques used in real-time fraud detection include:

4.3.1. Random Forests for Classification

Random Forests are one of the most used supervised learning models for fraud detection. This
ensemble method builds multiple decision trees from subsets of the training data and
combines their outputs to produce a final prediction. Random Forests are highly effective in
real-time fraud detection because they can process large datasets quickly and accurately. By
training on historical fraud data, these models learn to classify future transactions as
fraudulent or legitimate based on learned patterns. The model’s ability to handle high-
dimensional data makes it well-suited for complex fraud scenarios where a variety of

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 539

features—such as transaction amount, geographical location, and user behavior—need to be

considered simultaneously.

4.3.2. Autoencoders for Anomaly Detection

Autoencoders, a type of unsupervised learning model, are used for detecting anomalies in
data without requiring labeled examples of fraud. In fraud detection, autoencoders learn to
compress and reconstruct legitimate transaction data. If a new transaction deviates
significantly from the learned normal pattern, the autoencoder flags it as a potential fraud.
This approach is particularly useful in identifying emerging fraud patterns that might not be
represented in the training data. Autoencoders can analyze vast amounts of transaction data
in real time, identifying subtle anomalies that would otherwise go unnoticed in a traditional
rule-based system.

4.3.3. Recurrent Neural Networks (RNNs) for Sequential Data

Recurrent Neural Networks (RNNs) are effective at processing sequential data, making them
well-suited for detecting fraudulent patterns that occur over time. In real-time fraud detection,
RNNs can analyze sequences of transaction behaviors, such as multiple purchases in a short
period or unusual geographic shifts. By recognizing patterns in these sequences, RNNs can
predict whether a series of transactions is likely to be fraudulent. The ability of RNNs to retain
information from previous transactions allows them to detect long-term fraud strategies that
may unfold over multiple transactions.

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 540

Figure 2: Fraud Detection Using AI – Model Comparison

This figure compares the different machine learning models used in fraud detection. Random
Forests are employed for classification tasks, Autoencoders for anomaly detection,
and Recurrent Neural Networks (RNNs) for detecting sequential data patterns. Each model
contributes to detecting various types of fraudulent activities, improving the overall
robustness of fraud detection systems.

5. Case Study: AI-Driven Fraud Detection in Payment Systems

To illustrate the real-world application of AI-driven fraud detection, we present a case study
of a major financial institution that integrated machine learning models into its payment
systems to enhance fraud detection.

5.1. The Problem

The financial institution faced increasing challenges from sophisticated fraud attacks,
particularly in its credit card and online payment systems. Traditional rule-based detection
methods were generating a high number of false positives, causing disruptions to legitimate

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 541

customer transactions and leading to customer dissatisfaction. Additionally, the static nature
of the rules meant that the system struggled to keep up with rapidly evolving fraud tactics.

5.2. The Solution

To address these issues, the institution implemented a real-time fraud detection

system powered by a combination of Random Forests and Autoencoder models. The
Random Forest model was trained on historical transaction data, while the Autoencoder was
employed for anomaly detection in new, unfamiliar patterns. These models were integrated
into the institution’s transaction processing pipeline, allowing fraud to be detected within
milliseconds of the transaction occurring.

5.3. The Outcome

Within six months of deploying the new system, the financial institution saw a 35% reduction
in overall fraud losses. The number of false positives dropped by 50%, leading to fewer
customer service complaints and improved customer satisfaction. The AI-driven system was
able to detect emerging fraud patterns earlier, allowing the institution to respond more
quickly to new fraud tactics. This resulted in an 80% improvement in fraud detection
accuracy and a 20% reduction in operational costs associated with manual reviews.

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 542

Figure X: AI-Driven Fraud Detection Performance

This bar chart illustrates the improvements achieved through the AI-driven fraud detection
system, including a 35% reduction in fraud losses, a 50% reduction in false positives, and
an 80% improvement in detection accuracy.

6. Technical Challenges in Real-Time Machine Learning for Fraud Detection

While real-time machine learning offers significant advantages, it also introduces several
technical challenges that must be addressed to ensure its effectiveness.

6.1. Scalability and System Latency

Scaling real-time machine learning models to handle millions of transactions per second is a
major challenge for financial institutions. The models must be designed to process data with

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 543

minimal delay to avoid slowing down transaction processing times. This requires optimizing
both the machine learning algorithms and the underlying infrastructure. Solutions such
as distributed computing and cloud-based models can help reduce latency and improve the
system's ability to scale.

6.2. Balancing Precision and Recall

In fraud detection, there is often a trade-off between precision (the percentage of flagged
transactions that are actually fraudulent) and recall (the percentage of total fraud detected).
Too high of a focus on precision may result in missing actual fraud cases, while too high of a
recall may lead to an unacceptable number of false positives. Machine learning models must
be fine-tuned to strike the right balance between these two metrics. Techniques such
as threshold optimization and ensemble learning can help improve this balance.

6.3. Data Privacy and Security

As fraud detection models often rely on sensitive customer data, maintaining data privacy
and security is of utmost importance. Real-time machine learning models must comply with
regulatory requirements such as the General Data Protection Regulation
(GDPR) and California Consumer Privacy Act (CCPA). Techniques like differential
privacyand federated learning allow models to be trained on decentralized data sources
without sharing sensitive customer information, reducing the risk of data breaches.

7. Ethical Considerations

As AI systems play an increasingly central role in fraud detection, ethical considerations must
be at the forefront to ensure that these systems are both fair and accountable. While machine
learning models have the potential to reduce fraud at an unprecedented scale, their
widespread adoption brings concerns about bias, transparency, and accountability in
decision-making.

7.1 Bias in AI Models

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 544

One of the primary ethical concerns in AI-driven fraud detection is the potential for bias in
the models. Bias can emerge from the data used to train the models, leading to decisions that
disproportionately affect certain groups. For example, if historical fraud data
disproportionately reflects certain demographic groups or regions, the AI model may learn to
over-flag transactions from those groups as fraudulent. This can lead to discriminatory
outcomes, where legitimate transactions from certain demographics are more likely to be
flagged.

To mitigate this risk, financial institutions must implement bias detection techniques and
regularly audit their models to ensure they are not disproportionately targeting specific
groups. Techniques such as fairness constraints can be added to machine learning models to
reduce bias. These constraints ensure that models treat all demographic groups equally,
regardless of their representation in the training data.

7.2 Transparency and Explainability

Another ethical concern is the black-box nature of many machine learning models. Models
such as deep neural networks can make accurate predictions but provide little insight into
how those predictions are made. This lack of transparency can lead to mistrust among
customers, who may feel unfairly targeted by fraud detection systems. It also raises challenges
for regulatory compliance, as financial institutions must be able to explain how decisions are
made.

Explainable AI (XAI) offers a solution to this problem. By providing explanations for the
decisions made by machine learning models, XAI allows institutions to understand and justify
why certain transactions were flagged as fraudulent. This not only improves trust but also
ensures compliance with regulations that require transparency in automated decision-making
systems.

7.3 Accountability in AI-Driven Fraud Detection

Accountability is another crucial ethical consideration. In the event of false positives or missed
fraud cases, it is essential to determine where responsibility lies. Should the model developers,

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 545

data scientists, or the institution itself be held accountable for errors made by AI models?
Financial institutions must establish clear guidelines for the use of AI in fraud detection,
ensuring that human oversight is incorporated into the decision-making process.

Hybrid systems, where AI models flag suspicious transactions but human analysts make the
final decision, can help maintain accountability. This approach ensures that machine learning
models assist in the detection process, while humans remain responsible for critical decisions.

Figure 4: Ethical Considerations in AI Fraud Detection

8. Future Directions in AI-Driven Fraud Detection

The future of fraud detection lies in the continued evolution of machine learning models and
their integration with other emerging technologies. As fraudsters develop increasingly
sophisticated tactics, AI must also evolve to stay ahead of these threats.

8.1 Integration of Explainable AI (XAI)

As fraud detection models become more complex, the need for transparency and
explainability will only grow. In the future, Explainable AI (XAI) will likely become a
standard component of fraud detection systems. XAI will enable financial institutions to not

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 546

only detect fraud in real time but also provide clear, understandable explanations for why
specific transactions were flagged. This will improve customer trust and ensure compliance
with regulatory standards for transparency in AI-driven decision-making.

8.2 Blockchain for Enhanced Security

The integration of blockchain technology with AI-driven fraud detection holds significant
promise. Blockchain creates an immutable, decentralized ledger of transactions that can
enhance the security and traceability of financial transactions. When combined with AI,
blockchain can provide additional layers of security, making it harder for fraudsters to
manipulate transaction data. For example, AI can monitor blockchain-based transactions for
unusual activity, while blockchain ensures that transaction histories remain tamper-proof.

8.3 Hybrid AI Models

The next generation of fraud detection systems is likely to employ hybrid AI models that
combine supervised, unsupervised, and reinforcement learning techniques. By using a hybrid
approach, financial institutions can benefit from the strengths of each model type. For
example, supervised models can be used to identify known fraud patterns, while
unsupervised models detect anomalies in new, emerging fraud techniques. Reinforcement
learning can then be used to optimize detection strategies over time, improving both accuracy
and adaptability.

8.4 Real-Time Machine Learning at Scale

Scaling real-time machine learning systems to handle millions of transactions per second will
continue to be a focus for financial institutions. Advances in distributed computing, edge
computing, and cloud-based infrastructure will enable real-time models to process larger
datasets with lower latency. This will ensure that fraud detection systems remain effective as
transaction volumes grow in the digital economy.

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 547

8.5 AI for Predictive Fraud Detection

While current fraud detection systems are primarily reactive, future AI models may shift
toward predictive fraud detection. By analyzing transaction data, user behavior, and
historical patterns, predictive models will identify potential fraud risks before they occur.
These models could alert institutions to suspicious account activity, allowing them to take
preventive measures, such as flagging accounts for further review or temporarily halting
transactions until a threat is mitigated.

Figure 5: Future Directions in AI-Driven Fraud Detection

9. Conclusion

As fraudsters develop increasingly complex methods to deceive financial institutions, the role
of AI in fraud detection is more critical than ever. Real-time machine learning models have
revolutionized the way institutions detect and prevent fraud, enabling faster, more accurate
analysis of transaction data. By reducing false positives and improving detection rates, AI-
driven systems enhance both financial security and customer trust.

However, these advancements are not without challenges. Financial institutions must
navigate the technical complexities of scaling real-time machine learning systems while
ensuring data privacy and security. Additionally, ethical considerations such as bias,

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 548

transparency, and accountability must be addressed to ensure that AI-driven fraud detection
systems are fair and trustworthy.

The future of fraud detection will be shaped by the integration of technologies such
as Explainable AI and blockchain, as well as the development of hybrid AI models that
combine multiple machine learning techniques. By embracing these innovations, financial
institutions can stay ahead of fraudsters and protect their customers from increasingly
sophisticated threats.

References

1. Phua, A., Lee, V., Smith, K., & Gayler, R. (2010). A comprehensive survey of data
mining-based fraud detection research. Artificial Intelligence Review, 34(1), 1-14.
2. Zorarpacı, N., & Özel, S. (2016). A hybrid approach of support vector machine and
particle swarm optimization for real-time fraud detection in the credit card
industry. International Journal of Intelligent Systems and Applications, 8(12), 1-10.
3. Bahnsen, A., Aouada, D., Stojanovic, A., & Ottersten, B. (2016). Feature engineering
strategies for credit card fraud detection. Expert Systems with Applications, 51, 134-142.
4. Iqbal, N., Rehman, M. H., Jha, S. K., Hameed, A., & Al-Turjman, F. (2019). Real-time
machine learning-based system for financial fraud detection in large-scale
transactional data. IEEE Access, 7, 62209-62217.
5. Dal Pozzolo, M., Caelen, O., Bringay, Y. L., & Bontempi, P. (2015). Credit card fraud
detection and concept-drift adaptation with delayed supervised information. 2015
IEEE Symposium Series on Computational Intelligence, 803-810.
6. Arora, D., Varshney, R., & Gupta, R. (2017). A fraud detection system based on
machine learning and transaction analysis: A comparative study. Journal of Computer
Science and Information Technology, 5(2), 1-13.
7. Carcillo, F., Le Borgne, Y., Caelen, O., & Bontempi, G. (2018). Streaming active learning
strategies for real-life credit card fraud detection: Assessment and
visualization. International Journal of Data Science and Analytics, 5, 285–300.
8. Wu, M., Liu, X., & Liu, W. (2017). A scalable machine learning model for fraud
detection using a spark-based platform. Proceedings of the 2017 International Conference
on Data Science and Advanced Analytics (DSAA), 304-313.

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.
Journal of Artificial Intelligence Research and Applications
By Scientific Research Center, London 549

9. Bhattacharyya, J., Ghosh, S., & Bose, A. (2008). An unsupervised learning approach to
real-time credit card fraud detection. Proceedings of the 19th International Conference on
Pattern Recognition (ICPR), 1183-1187.
10. Moslemi, R., & Hashemi, S. H. (2019). A novel real-time hybrid approach for credit
card fraud detection. Computers & Security, 84, 349-362.
11. Bolton, S., & Hand, D. (2002). Statistical fraud detection: A review. Statistical Science,
17(3), 235-249.
12. Vlassis, S., & Likas, A. (2002). A greedy EM algorithm for Gaussian mixture
models. Neural Processing Letters, 15, 77-87.
13. Dal Pozzolo, A., Caelen, O., Johnson, R., Waterschoot, S., & Bontempi, G. (2015).
Calibrating probability with undersampling for unbalanced classification. 2015 IEEE
Symposium Series on Computational Intelligence, 159-166.
14. Jurgovsky, A., Granitzer, M., & Ziegler, J. (2018). Sequence classification for credit-card
fraud detection. Expert Systems with Applications, 100, 234-245.
15. Whitrow, M., Hand, P., Juszczak, I., Weston, D., & Adams, D. (2009). Transaction
aggregation as a strategy for credit card fraud detection. Data Mining and Knowledge
Discovery, 18(1), 30-55.
16. Cortes, C., Jackel, L. D., & Denker, W. S. (1991). Learning algorithms for pattern
classification with the general regression neural network. Pattern Recognition, 24(12),
1149-1157.
17. Sahin, M., & Duman, E. (2011). Detecting credit card fraud by decision trees and
support vector machines. Proceedings of the 2011 International MultiConference of
Engineers and Computer Scientists (IMECS), 442-447.
18. Hand, D. J., Blunt, G., Kelly, M. G., & Adams, N. M. (2005). Data mining for fun and
profit: Tackling the unbalanced classification problem in fraud detection. Journal of
Royal Statistical Society, 66(3), 321-331.

Journal of Artificial Intelligence Research and Applications

Volume 3 Issue 2
Semi Annual Edition | Jul - Dec, 2023
This work is licensed under CC BY-NC-SA 4.0.