Abstract Review-01

Under Esteemed Guidance of

Mrs. Swarna Ramya
Assistant Professor Submitted By

M SANDEEP

(20KT1A0597)
CREDIT CARD DETECTION
USING RANDOM
FORESTALGORITHM
ABSTRACT

In This paper we mainly focus on credit card fraud detection in real world. Here
the credit card fraud detection is based on fraudulent transactions. Generally credit
card fraud activities can happen in both on line and offline. But in today's world
online fraud transaction activities are increasing day by day. So in order to find the
online fraud
transactionsvariousmethodshavebeenusedinexistingsystem.Inproposedsystemwe
use Random Forest Algorithm (RFA) for finding the fraudulent transactions and
the accuracy of those transactions.

This algorithm is based on supervised learning algorithm where it uses

decision trees for classification of the dataset. After classification of dataset a
confusion matrix is obtained. The performance of Random Forest Algorithm is
evaluated based on the confusion matrix. The results obtained from processing
the dataset.
Existing System

InExisting system, Research on a case study including the identification of credit card fraud in the
current system,where data normalization is used prior to cluster analysis and with the outcomes
received from the application of artificial neural networks and cluster analysis to fraud detection
has demonstrated that neuronal inputs can be reduced bygrouping characteristics. Additionally,
employing normalized data and MLP training on the data can yield excellent results. The
foundation of this study was unsupervised learning. This paper was significant since it identified
newtechniques for detecting fraud and improved the accuracy of the findings.The paper's data set
is derived from actual transactional data collected by a major European corporation; all personal
information is kept private. An algorithm's accuracy is about 50%.
Proposed System

In proposed System, we are applying random forest algorithm for

classification of the credit card dataset. Random Forest is an algorithm for
classification and regression. Summarily, it is a collection of decision tree
classifiers. Randomforesthas advantage overdecision tree as itcorrects the habit
of over fitting to their training set. A subset of the training set is sampled
randomly so that to train each individualtreeand thenadecisiontreeisbuilt,each
node then splits on a feature selected from a random subset of the full feature
set. Even for large data sets with many features and data instances training is
extremely fast in random forest and because each tree is trained independently
of the others. The Random Forest algorithm has been found to provide a good
estimate of the generalization error and to be resistant to over fitting.
Requirements

The technical requirements for the software products are outlined in the
requirements specification. It is the initial phase ofthe procedure for requirements
analysis .
performance,andsecurityrequirementsofacertainsoftwaresystem.Asoftwarerequ
irementsspecification serves the purpose of giving a thorough overview of the
software project, including its parameters and objectives.
 HARDWARE REQUIREMENTS

Intel Processor
4gb Ram
260 gb Hard
Disk

SOFTWARE
REQUIREMEN
 ALGORITHM

RANDOM FOREST ALGORITHM

Random forests is a supervised learning algorithm. It can be used both for

classification and regression. It is also the most flexible and easy to use
algorithm. A forest is comprised of trees. It is said that the more trees it
has, the more robust a forest is. Random forests create decision trees on
randomly selected data samples, gets prediction from each tree and selects
the best solution by means of voting. It also provides a pretty good
indicator of the feature importance.
IMPLEMENTATION OF ALGORITHM – (RFA)

1. Data Collection
2. Data Preprocessing
3. Feature Selection
4. Train/Test Split
5. Model Training
6. Model Evaluation
7. Hyperparameter Tuning
SYSTEM ARCHITECTURE
MODULE DESCRIPTION
CONCLUSION

The Random Forest algorithm will perform better with a larger number of training
data, but speed during testing and application will suffer. Application of more
preprocessing techniques would also help. The SVM algorithm still suffers from the
imbalanced dataset problem and requires more preprocessing to give better results at
the results shown by SVM is great but it could have been better if more
preprocessing have been done on the data. Random Forest Algorithm in credit card
fraud detection system and the final optimization results indicates the optimal
accuracy for Random Forest Algorithm is 98.6%. Although random forest obtains
good results on given data set, there are still some problems such as imbalanced data.
Our future work will focus on solving these problems.
FUTURE SCOPE

It is evident from the above review that several machine learning algorithms are used
to detect fraud, but the findings are not satisfactory. As a result, we'd like to use deep
learning algorithms to reliably detect credit card fraud.