Abstract-Intrusion detection plays important role in examining the malicious activity which occurs in a network or a system. Its a software or a device which scans the system or network for a abnormal activity. Since the connectivity between computers have grown up in large numbers the need for intrusion detection plays a vital role in network security. Even though various machine learning techniques and statistical methodologies have been used to build different types of Intrusion Detection Systems the perfromance mainly depends on the accuracy. Reducing the false positives and increasing efficiency of the detection rate are the main challenges in Intrusion detection system. The recent works have utilized various technologies to improve the false positives and detection rate.But a more intelligent approach will be to make use of large network traffic data and an incremental classification methodology to overcome this issue.The proposed approach makes use of various Machine learning techniques like Logistic Regression , Gaussiann Naive Bayes, DecisionTree Classifier,Random Forest Classifier, Gradient Boosting Classifier and Support Vector Machine (SVM) . KDD dataset is used to perform the classification, and its observed that Random Forest Classifier performed well compared to others, meeting the demands of the network intrusion. Keywords IDS, KDD dataset, SVM,DoS

• Introduction

The main scope of this project is to develop an network intrusion detection system which can distinguish the normal connections from bad connections. Installation of IDS in a network system is very important in order to analyse the quantity and type of cyber attacks. The analysis will the help the organisations to change their security systems or implement more effective controls.It can also help the organisations to identify bugs or problems within their network device configurations. Hence its very important to deisgn a user friendly and effective Intrusion detection System. It can be used in government organisations where the sensitive data is stored, like passport systems, social Id card system, banks, military system or even the social media platforms like facebook, twitter can also use them in order to protect the intruders from stealing the sensitive data.Hence we are aimed at creating graphics user interface application which can detect the network intrusion using the machine learning model with low false positives , high accuracy and also can summarise the types of atacks occuring over the network.

• Background and Related Work

In this section, we will analyse the past research works on the Intrusion Detection System.[1] A.A.diro and chilakurti proposed a distributed attack detection scheme using deep learning. The results were found to be 96% accurate, but the system lacked well built for visualizing the results.. Hoda et al. [2] Hodo et al. Proposed a neural network based IDS for low-capacity devices. The experimental results of the work