Cyber Fraud App Detection Using Machine Learning
ABSTRACT
As of late, the uses of advanced mobile phones are expanding relentlessly and furthermore
development of Android application clients are expanding. Because of development of
Android application client, some gate crashers are making vindictive android application as
instrument to take the delicate information and data for fraud and misrepresentation portable
bank, versatile wallets. There are such a large number of malevolent applications discovery
instruments and programming’s are accessible. Be that as it may, a viably and productively
vindictive application recognition device expected to handle and deal with new complex
pernicious applications made by interloper or programmers. This paper Utilizing Machine
Learning approaches for distinguishing the malignant android application. First, dataset of
past pernicious applications has to be obtained with the assistance of Help vector machine
calculation and choice tree calculation make up correlation with preparing dataset. The
prepared dataset can foresee the malware android applications up to 93.2 % obscure/new
malware portable application..
The main steps performed through this framework are sketched as follows:
1. A set of features is computed for every binary file in the training or test datasets, based on
many possible ways of analysing a malware.
2. A Deep learning system based firstly on one-sided perceptron’s, and then on feature
mapped one-sided perceptron’s and a kernelized one-sided perceptron’s, combined with
feature selection based on the F1 and F2 scores, is trained on a medium-size dataset
consisting of clean and malware files. Cross-validation is then performed in order to choose
the right values for parameters. Finally, tests are performed on another, non-related dataset.
The obtained results were very encouraging.
3. In the end we will analyse different aspects involved in the scale-up of our framework to
identifying malware files on very large training datasets.
EXISTING SYSTEM:
Suleiman et al. [18] proposed a classification approach based on parallel machine
learning to detect Android malware. Depends on real malware samples and benign
applications have derived from it, a total of 179 training features were extracted and
divided into API calls and commands related: 54 features. App permissions: 125. A
composite classification model was developed from a parallel set of heterogeneous
classifiers, namely Simple Logistic, Naïve Bayes, Decision Tree, PART, and RIDOR.
DISADVANTAGES OF EXISTING SYSTEM:
Accuracy of these Models are less and prediction of results are not
correct
All types of features are not taken as input.
Algorithm detects genuine or fake only.
PROPOSED SYSTEM:
The schematic representation of the architecture of this work shows the first
phase starting with the mixture of malware and benign Android Application
Package (APK) files taken from a dataset that is publicly available to everyone.
Then, we extract the static features by our own Python script written in Jupiter
notebook environment. The features are API-calls and permissions. These
features are formatted and stored in Comma Separated Values (CSV) file as a
data frame that serves the training process. Finally, we test all classifiers using
10-folds cross-validation, in addition to calculating the appropriate metrics for
the evaluation
ADVANTAGES OF PROPOSED SYSTEM:
Accuracy of the model is high deep learning methods are used to train model
Prediction of malicious or not is also possible form this model.
Time taken for prediction is less.
SYSTEM REQUIREMENTS
HARDWARE REQUIREMENTS:
• System : Intel(R) Core(TM) i3-7020U CPU @ 2.30GHz
• Hard Disk : 1 TB.
• Input Devices : Keyboard, Mouse
• Ram : 4 GB.
SOFTWARE REQUIREMENTS:
Operating system : Windows XP/7/10.
Coding Language : Python
Tool : Anaconda
Interface : flask
Database : Mysql
