Scribd Logo
Upload

Welcome to Scribd!

  • 67 views

    Synopsis FinalFINAL

    Uploaded by

    Hanu M
    This project aims to develop a machine learning model to identify fraudulent online transactions. A team of 4 students will collect transaction data, preprocess it, extract features, and train models like fully connected neural networks and XGBoost to classify transactions as fraudulent or legitimate. These models can help detect fraud with high accuracy, enhancing security for online payments and preventing losses. Traditional methods are limited, while machine learning can analyze complex transaction patterns in large datasets.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    Synopsis FinalFINAL

    Uploaded by

    Hanu M
    67 views4 pages
    This project aims to develop a machine learning model to identify fraudulent online transactions. A team of 4 students will collect transaction data, preprocess it, extract features, and train models like fully connected neural networks and XGBoost to classify transactions as fraudulent or legitimate. These models can help detect fraud with high accuracy, enhancing security for online payments and preventing losses. Traditional methods are limited, while machine learning can analyze complex transaction patterns in large datasets.

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This project aims to develop a machine learning model to identify fraudulent online transactions. A team of 4 students will collect transaction data, preprocess it, extract features, and train models like fully connected neural networks and XGBoost to classify transactions as fraudulent or legitimate. These models can help detect fraud with high accuracy, enhancing security for online payments and preventing losses. Traditional methods are limited, while machine learning can analyze complex transaction patterns in large datasets.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    67 views4 pages

    Synopsis FinalFINAL

    Uploaded by

    Hanu M
    This project aims to develop a machine learning model to identify fraudulent online transactions. A team of 4 students will collect transaction data, preprocess it, extract features, and train models like fully connected neural networks and XGBoost to classify transactions as fraudulent or legitimate. These models can help detect fraud with high accuracy, enhancing security for online payments and preventing losses. Traditional methods are limited, while machine learning can analyze complex transaction patterns in large datasets.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    You are on page 1of 4

    PROJECT SYNOPSIS

    Project Title
    ONLINE TRANSACTION FRAUD DETECTION USING MACHINE LEARNING

    TEAM MEMBERS
    Girish Kumar T P Jayanth Gowda B R Keerthan B K Mohan Bevinatti K S

    ABSTRACT
    Now a days Digital transactions are rapidly increasing as it results in increasing
    online payment frauds too. In fact, according to the Reserve Bank of India, comparing
    March 2022 to March 2019, digital payments have risen in volume and value by
    216% and 10%, respectively. People are starting to go all-in with digital transactions,
    but one can’t deny the security issues that loom, and know-how when it comes to
    online payments. Few years ago, we could have barely seen the online payment, but
    today UPI payment QR code installed at doorstep. This invited the hoaxers and
    attackers to develop fraudulent transactions and fool people for some amount of
    money. Fortunately, the online transactions are monitored and hence could be
    analyzed using the latest tools. In this system, an attempt is made to develop a
    machine learning model to identify fraudulent transactions in a transaction’s dataset.

    INTRODUCTION

    Mobile payment has gained significant popularity as a mainstream payment method,


    leading to a high volume of transactions on online trading platforms. Unfortunately,
    this popularity also attracts criminals who exploit the complex network environment to
    commit fraud. Such fraudulent activities not only harm consumers but also impede the
    healthy growth of the online economy. Consequently, effective transaction fraud
    detection becomes a vital tool in combating network transaction fraud. Traditional
    fraud detection approaches primarily rely on statistical and multi-dimensional analysis
    techniques. However, these verification based methods struggle to uncover the
    underlying patterns in transaction data, limiting their effectiveness. On the other hand,
    big data technology and machine learning algorithms offer efficient solutions for
    detecting transaction fraud. Machine learning, particularly when applied to large
    datasets, can capture important features that traditional statistical methods fail to
    describe. By utilizing suitable machine learning techniques, we can build models
    based on existing transaction data to detect network transaction fraud, thereby
    mitigating associated losses. In 2018, Zhaohui Zhang proposed a reconstructed feature
    convolutional neural network prediction model specifically tailored for transaction
    fraud detection. This model demonstrated improved stability and classification
    effectiveness compared to other convolutional neural network models. However, a
    challenge remains in achieving high detection accuracy due to imbalanced sample
    labels. To address this, the paper introduces two fraud detection algorithms: one based
    on a Fully Connected Neural Network and another utilizing XGBoost. The former
    algorithm integrates two neural network models with different cross-entropy loss
    functions, enabling a quick and convenient design process for the combined model.
    The latter algorithm leverages Hyperopt to optimize the XGBoost classifier, resulting
    in a fraud detection model with superior performance by selecting the best parameters.
    These two algorithms serve different application scenarios.
    SYSTEM ARCHITECTURE

    METHODOLOGY

    1. Data Collection
    2. Data preprocessing
    3. Feature Engineering
    4. Splitting the Data
    5. Choosing a model
    6. Training the model
    7. Evaluation
    8. Hyper parameter tuning
    9. Validation
    ADVANTAGES

     Highly security from unauthorized use of UPI.


     Avoid fraud usage of card through online transactions.
     Detect if card used by others if card lost.
     High accurate results.

    REFERENCES

    [1]. Mohammed, Emad, and Behrouz Far.“Supervised Machine Learning Algorithms


    for Credit Card Fraudulent Transaction Detection A Comparative Study.” IEEE
    Annals of the History of Computing, IEEE, 1 July 2018,
    doi.ieeecomputersociety.org/10.1109/IRI.2018.00025.
    [2]. Randhawa, Kuldeep, et al. “Credit Card Fraud Detection Using AdaBoost and
    Majority Voting.” IEEE Access, vol. 6, 2018, pp. 14277–14284.,
    doi:10.1109/access.2018.2806420.
    doi:10.1109/colcomcon.2017.8088206.
    [3]. http://www.rbi.org.in/Circular/CreditCard
    [4].https://www.ftc.gov/news-events/pressreleases/2019/02/imposter-scams-
    topcomplaints-made-ftc-201

    You might also like