Scribd
Upload
53 views10 pages

Credit Card Fraud Detection

The document discusses credit card fraud detection. It describes the problem of identifying fraudulent transactions from past transaction data. The aim is to detect 100% of fraudulent transactions while minimizing incorrect classifications. Features like user spending patterns and locations are used. The testing process and results achieving 94.49% accuracy are also summarized.

Uploaded by

vivekvivek05361
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
53 views10 pages

Credit Card Fraud Detection

The document discusses credit card fraud detection. It describes the problem of identifying fraudulent transactions from past transaction data. The aim is to detect 100% of fraudulent transactions while minimizing incorrect classifications. Features like user spending patterns and locations are used. The testing process and results achieving 94.49% accuracy are also summarized.

Uploaded by

vivekvivek05361
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 10

Credit Card

Fraud Detection

Guided By Mr.Vishnu Prakash


Sharma
Problem Statement
The CreditCard Fraud Detection Problem includes modelingpast credit card
transactions with the knowledge of the ones that turned out to be fraud. This
model is then used to identify whether a new transaction is fraudulent or not.
Our aim here is to detect 100% of the fraudulent transactions while
minimizing the incorrect fraud classifications.
The credit card fraud detection features use user behaviour and location
scanning to check for unusual patterns. These patterns include user
characteristics such as user spending patterns as well as usual user
geographic locationsto verify his identity. If any unusualpattern is detected,
the system requires revivification.
Frauds over the
years
Testing Process

. We choose Credit Approval Dataset due to its contents i.e. it contains about 15
features divided into 3 classes (integer, real and categorical) which would surely
help in training the model to achieve a high accuracy. This dataset was very
interesting because there was a good mix of attributes -- continuous, nominal
with small numbers of values, and nominal with larger numbers of values. There
were also a few missing values. For building the neural network used keras with
TensorFlow as a backend to build the network.
The network is a 3-layer network with 11 input nodes 3 nodes in middle layer
and one node in Output layer. Visualization of the results is done using
Matplotlib
Project Timeline

Week 1 Week 2 Week 3 Week 4 Week 5

Deployed
Task Choosing project Gathering data Testing model Frontend Coding
backend on
name
django

Studying Enhancing algo


Choosing Frontend design
research paper for optimization Backend
Task libararies
name connected to
local server
Setting Prediction Validating the
Environme Trainnig data of result web application
Task nt
Final Outlook
name
Results &
Findings
Our model achieved a accuracy of 94.49 % for fraud detection and the affected population (which
were considered fraud but were not fraud was found to be 22.60 %. The results mayvary if you
run the same jupyternotebook because initialization of the weightsof the nodes of SOM grid is
done by randomly selecting the records/ patterns from the input space i.e. randomly selecting the
records from the given dataset. Since,we have done training for 100 iterations and weights are
randomly initialized every time, convergence may vary. We may try with different iterations like
100, 150, 200 etc. to have betterconvergence. You may also store the weightsof the SOM for
which you achievebetter accuracy.
Conclusion

From the above comparative analysis of the various creditcard fraud detection
techniques it is clear that Artificial Neural Networks performs best in this
scenario. But the drawbacks of Artificial Neural Networks is that they are very
expensive to train and can be easily over trained. In order to minimize their
expense we need to create a hybrid of neural network with some optimisation
technique. Optimisation techniques that could be successfully paired with
Neural Networkare Genetic Algorithm, Artificial Immune System,Case Based
Reasoningand any other similar optimisation technique
Resources
[1] Dataset:
http://archive.ics.uci.edu/ml/data
sets/Credit+Approval

IEEE Technical Paper:


https://www.researchgate.net/publication
/334761474_Real-
time_Credit_Card_Fraud_Detection_Using_Machine_Learning

[4]Website:
https://towardsdatascience.com/credit-card-fraud-
detection- using-self-organizing- featuremaps-
f6e8bca707bd
Does Anyone
Have a
Questions?

You might also like