Maximizing Impact_ Nonprofit Website Planning, Budgeting, and Design.pdf
Android Malware
1. Under the guidance of
Mrs.Thamizhisai
Assistant Professor
Department of CSE
RAAKCET.
DETECTION OF ANDROID MALWARES USING
RECURRENT NEURAL NETWORKS
TEAM MEMBERS REGISTER NUMBER
K.Kaviarasan 19TD1515
P.Nambiraju 19TD1521
B.Praisen 19TD1524
A.Praveen kumar 19TD1525
2. 2
OBJECTIVE OF THE PROJECT
Title: Detection of Android Malwares using RNN
The main goal of this project is to develop an efficient
deep learning model to detect the android malwares from
the genuine files.
3. 3
DOMAIN OF THE PROJECT
Domain: Deep Learning
Explanation: Deep learning is a class of machine learning
algorithms that uses multiple layers to progressively
extract higher-level features from the raw input.
4. 4
DOMAIN OF THE PROJECT
It is extremely beneficial to data scientists
Deep learning makes this process faster and easier.
Deep Learning can apply to the complex problems
6. EXPLANATION
Preprocessing
Data preprocessing is required tasks for cleaning the data and making it
suitable which also increases the efficiency of a machine/deep learning
model.
Getting the dataset
Importing libraries
Importing datasets
Finding Missing Data
Encoding Categorical Data
Normalization
Normalization is to change the values of numeric columns in the dataset to
use a common scale.
7. EXPLANATION
Splitting the Dataset into the Training set and Test set
In machine/deep learning data preprocessing, we divide our dataset into a
training set and test set.
Training Set: A subset of dataset to train the machine learning model, and
we already know the output.
Test set: A subset of dataset to test the machine learning model, and by
using the test set, model predicts the output.
8. 8
ALGORITHM&TECHNIQUES
FCBF (Fast correlation based filter) is a multivariate feature selection
algorithm used in proposed system
A recurrent neural network (RNN) is a class of artificial neural
networks used in proposed system
Models under the Recurrent Neural Network are:
Long Short Term Memory (LSTM)
Gated Recurrent Unit (GRU)
9. 9
ALGORITHM&TECHNIQUES
LONG SHORT TERM MEMORY (LSTM)
LSTM is a kind of recurrent neural network (RNN) design applied in
deep learning field.
LSTM has feedback connection that is unrelated to standard feed
forward neural networks.
10. 10
ALGORITHM&TECHNIQUES
GATED RECURRENT UNIT (GRU)
GRU is a type of deep learning algorithm that is enhanced from the
LSTM algorithm to minimize the complication of the algorithm by
using update gate and reset gate.
The update gate is used to regulate hidden state volume to be
forwarded to the next state.
11. 11
ADVANTAGES OF PROPOSED SYSTEM
The project is expected to show better results by implementing
Recurrent Neural Networks (RNN) and FCBF.
It Improves the Runtime and provide better result than existing
system.
12. 12
ADVANTAGES OF PROPOSED SYSTEM
In this existing model, a fast Android malware detection
framework (FAMD), which slows the runtime of the
software.
The proposed method can achieve an accuracy of 99%
and an F1-score of 98%.
13. HYPERPARAMETER SETTINGS
Optimizer
The role of the optimizer is to improve the accuracy of the model by
reducing the error rate.
Eg. Adam, SGD, Adagrad
Dropout
Dropout are generally used to reduce the complexity between the links
present in the fully connected dense layer.
It is a user dependent variable which take the input value from 0 to 1.
Epoch
The training method is repeated number of times is known as the epoch.
Depends on the training data, the number of epochs are considered.
14. HYPERPARAMETER SETTINGS
LSTM and GRU Units
The number of LSTM and GRU units is the number of memory cells in the
LSTM and GRU network.
It denotes the capable of remembering the facts and matches it with the past
evidence.
The information in the memory units are moved further in the following
time step for the further training.
Max Pooling
Max pooling is a pooling operation that selects the maximum element from
the region of the feature map covered by the filter. Thus, the output after
max-pooling layer would be a feature map containing the most prominent
features of the previous feature map.
15. HYPERPARAMETER SETTINGS
Activation Function
At the output layer, an activation function is used to decide the probability
of the message.
Eg. ReLU, Tanh, Sigmoid.
Learning Rate
Learning rate is a tuning parameter in an optimization algorithm that
determines the step size at each iteration while moving toward a minimum
of a loss function.