Intrusion Detection System Framework Using Stack auto encoder and GRU in the cloud Environment

Abstract

Cloud usage has increased and protecting the data in cloud environment is very important. Many intruders attack the cloud environment. In this paper, a novel Deep learning framework has been proposed to detect intrusion in the cloud environment using Deep Learning based Gated Recurrent Unit (GRU). The input dataset is first pre-processed with data cleaning, and normalization techniques applied to remove unwanted data so that data quality is improved. A Stack based auto encoder is then used to extract features from the pre-processed data, reducing dimensionality and emphasizing important characteristics. Afterwards, the Mountain Gazelle Optimizer (MGO) is employed to select the most relevant features. Then it is classified by GRU. Different kinds of attack are found and normal data and abnormal data are classified. The dataset KDDCup-99 and CIC-IDS 2017 are used. It gives better accuracy and reduced false positive rate.

Specification

Description:The following specification particularly describes the nature of the invention and the manner in which it is performed:
FIELD OF THE INVENTION
 Protecting the security of cloud actions has become crucial as cloud activities have increased.
 Intrusion detection systems (IDS) play a vital role in protecting cloud infrastructures against unwanted access and harmful activity.
 This study provides an outlandish strategy to intrusion detection that is specifically designed for cloud systems, utilizing deep learning techniques.
BACKGROUND OF THE INVENTION
Cloud computing provides numerous advantages, including flexibility, cost-effectiveness, scalability, and accessibility. Based on demands, users may quickly scale up or down resources and pay only for what they use. IDS collects and examines data from various places, including system logs, network packets and application activities, applying techniques such as anomaly detection, signature-based detection and heuristic analysis. When a potential intrusion or security incident is found, the IDS triggers automated responses or generates alerts such as blocking malicious traffic, notifying security personnel, or logging the event for further investigation.
IDS is a crucial element of modern cybersecurity frameworks, designed to monitor and analyses cloud system activity showing malicious behaviour, unauthorized access, and policy violations.
Autoencoders are used in IDS because they can extract and represent key aspects from complicated network data. Autoencoders can successfully capture underlying patterns and abnormalities in network traffic by compressing input data and recreating it.
Algorithm like (ACST) and (ANFIS) Classifier, (COA-GS), (GMM) and K-Means clustering algorithms are used with the RF, CNN-based inception, SIGMOD and DAERF are used. Their performance are not as expected. The proposed used stack auto encoder for feature extraction and GRU for classification. Mountain Gazelle Optimizer (MGO) is used for optimization and so it predicts accurately.
OBJECTIVE OF THE INVENTION
 The goal is to develop a deep learning model for intruder's detection in cloud environment, to potentially replace the updatable supervised machine learning classification models by predicting results in the form of best accuracy by comparing supervised algorithm
 A Model is proposed to detect the attack in the cloud environment
 Stack auto encoder is used for feature detection and GRU used for classification of normal and
3
abnormal traffic.
SUMMARY OF INVENTION
 Deep learning is the area in information technology which pacts with the usage of certain algorithms for automating certain tasks which are very monotonous and can be done easily. It is highly helpful in the prediction of certain thoughts and classification problems, quick trading prediction. Here in this work the deep learning algorithm is used for prediction and detection of attack in cloud environment. The analytical approach included data cleaning and processing, missing value analysis, exploratory analysis, and model construction and evaluation. The highest accuracy score, as well as the best accuracy on a public test set, will be discovered. Our analysis provides a comprehensive guide to sensitivity analysis of model parameters with regard to performance in detecting the intruders in cloud environment by accuracy calculation. , Claims:We Claim:
1. This Invention proposed a Model is designed to detect the intruders attack in cloud environment.
2. Designed model is made to spot the abnormal data. From data set multi-variant analysis is done.
3. Model is designed to detect the abnormal data or attacks by accuracy calculation

Documents