Work

Identifying inappropriate access points using machine learning algorithms RandomForest and KNN

Year: 2023
Type: dissertation
Abstract: A wireless network is being used by a wide range of users in today’s world from large organizations to schools to government offices adding to this WiFi hotspots are available at many public places which are used by thousands of people. The wireless network of an organization may facilitate the transfer of highly important and sensitive information and even normal person using a wireless network may transfer their sensitive information through the network. The transfer of large amounts of data has attracted the attention of people who might be looking to steal or leak the sensitive information that is being transferred through a wireless network and these people are able do this by configuring access points that are inappropriate. People may unknowingly connect to these inappropriate access points and may result in the attacker stealing valuable information which may lead to huge financial and personal losses. So, it is important to identify the presence of unauthorized access points in a wireless network and a system is proposed here to do so. The system proposed here will use machine learning classifiers for identifying the presence of unauthorized access points in a wireless network. The system will be developed based on the Aegean WiFi Intrusion Dataset (AWID) intrusion detection dataset. The data will be pre-processed and the important features from the data will be selected using the Analysis of variance (ANOVA) technique. These features will be used for training the machine learning classifiers K-nearest neighbor (KNN), Random Forest, Support Vector Machine (SVM), Genetic Algorithm (GA). The accuracy and precision of the machine learning classifiers will be found out for evaluating the effectiveness of the classifiers and the classifier with best performance will be used for creating desktop application that is able to identify the networks with inappropriate access points based on the network features provided as input to it. It found as the result of the approach that the machine learning classifiers are able to effectively identify wireless networks with inappropriate access points.
Source: NORMA
Author: Vaidya, Tushar Sanjaykumar
Language: English
Open Access status: green
OpenAlex ID: W7038490907
Cites: 21
Cited by: 3
Related to: 10