Loading...
Location
Virtual - Microsoft Teams Meeting
Start Date
2-7-2020 1:00 PM
End Date
2-7-2020 1:30 PM
Timezone
Ethiopian Time
Start Date (EST)
2-7-2020 6:00 AM
End Date (EST)
2-7-2020 6:30 AM
Description
Most researches have been conducted to develop models, algorithms and systems to detect intrusions. However, they are not plausible as intruders began to attack systems by masking their features. While researches continued to various techniques to overcome these challenges, little attention was given to use data mining techniques, for development of intrusion detection. Recently there has been much interest in applying data mining to computer network intrusion detection, specifically as intruders began to cheat by masking some detection features to attack systems. This work is an attempt to propose a model that works based on semi-supervised collective classification algorithm. For this study, data mining algorithms were first selected based on efficiency and accessibility criteria. An experiment was conducted using real .arff dataset to develop the model. The result shows that meta.Filtered Collective Classifier is appropriate to detect intrusions with hidden features, which scored the best classification accuracy of 96.2%.
PDF Copy
Included in
Computational Engineering Commons, Computer and Systems Architecture Commons, Digital Communications and Networking Commons, Other Computer Engineering Commons, Robotics Commons
Efficient Data Mining Algorithm Network Intrusion Detection System for Masked Feature Intrusions
Virtual - Microsoft Teams Meeting
Most researches have been conducted to develop models, algorithms and systems to detect intrusions. However, they are not plausible as intruders began to attack systems by masking their features. While researches continued to various techniques to overcome these challenges, little attention was given to use data mining techniques, for development of intrusion detection. Recently there has been much interest in applying data mining to computer network intrusion detection, specifically as intruders began to cheat by masking some detection features to attack systems. This work is an attempt to propose a model that works based on semi-supervised collective classification algorithm. For this study, data mining algorithms were first selected based on efficiency and accessibility criteria. An experiment was conducted using real .arff dataset to develop the model. The result shows that meta.Filtered Collective Classifier is appropriate to detect intrusions with hidden features, which scored the best classification accuracy of 96.2%.
