Performance Evaluation of Intrusion Detection using Linear Regression with K Nearest Neighbor
Journal Title: International Journal of Trend in Scientific Research and Development - Year 2019, Vol 4, Issue 1
Abstract
Starting late, the colossal proportions of data and its unfaltering augmentation have changed the essentialness of information security and data examination systems for Big Data. Interference acknowledgment structure IDS is a system that screens and analyzes data to perceive any break in the structure or framework. High volume, arrangement and quick of data made in the framework have made the data examination strategy to perceive ambushes by ordinary strategies problematic. Gigantic Data frameworks are used in IDS to oversee Big Data for exact and profitable data examination process. This work introduced Regression based gathering model for interference area. In this model, we have used direct backslide for feature decision examination, and built an interference revelation appear by using Na¯ve bayes classifier on concern organize. Presently used KDD99 to plan and test the model. In the examination, we displayed an assessment between LRKNN Linear Regression based K Nearest Neighbor and CM KLOGR Confusion Matrix based Kernel Logistic Regression classifier. The eventual outcomes of the assessment exhibited that LRKNN show has unrivaled, decreases the planning time and is viable for Big Data Content mining based IDS can beneficially perceive obstructions. Linear Regression based K Nearest Neighbor LRKNN is one of the progressing overhauls of chaste knn computation. LRKNN deals with the issue of self governance by averaging all models made by ordinary one dependence estimator and is suitable for relentless learning. This way of thinking is sharp framework interference acknowledgment system using LRKNN estimation for the recognizable proof of different sorts of attacks. To evaluate the execution of our proposed system, we drove tests NSL KDD enlightening list. Trial results make evident that proposed model dependent on LRKNN is profitable with low FAR and high DR. Deepa Hindoliya | Prof. Avinash Sharma "Performance Evaluation of Intrusion Detection using Linear Regression with K Nearest Neighbor" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-4 | Issue-1 , December 2019, URL: https://www.ijtsrd.com/papers/ijtsrd29525.pdf Paper URL: https://www.ijtsrd.com/computer-science/other/29525/performance-evaluation-of-intrusion-detection-using-linear-regression-with-k-nearest-neighbor/deepa-hindoliya
Authors and Affiliations
Deepa Hindoliya | Prof. Avinash Sharma
Keywords
Related Articles
- EP ID EP685682
- DOI -
- Views 154
- Downloads 0
To Compare The Effect Of Proprioceptive Neuromuscular Facilitation and Static Stretching on Flexibility of Hamstring Muscle: A Comparative Study
Background: Flexibility is an important physiological component of physical fitness and reduced flexibility can cause inefficiency in the workplace and is also a risk factor for low back pain. Increasing hamstring flexib...
Air Water System Design using Revit Mep for a Residential Building
In this project we discussed the study and performance of air conditioner, air refrigeration and water conditioner system in a single unit. The main objective of this project is to develop the multifunctional system whic...
The Relationship of Coping Mechanisms to the Role of 4PS
This study analyzed the relationship between the roles and coping mechanisms of the recipient of 4Ps in the province of Northern Samar considering the 4Ps recipients as the representatives of the poor families. It descri...
Regional Determinants of Central Asian Development
Contemporary complex international challenges demonstrates inability for many nations to solve a large range of issues of internal development solely or independently. This situation has a many aspects and the search for...
Database for Mobile Application
As we know that database is a collection of interrelated data i.e. it is composed of collection of files that are linked in such a way that information from one of files may be combined with information from other files...