Informatics Publishing Limited

Anurag Lal ¹, Vivek Dubey ²

Affiliations
1 Computer Science & Engineering, CIT, Rajnandgaon, IN
2 Computer Science & Engineering, SSCET, Bhilai, IN

Abstract

In this paper loss of packets in TCP is detected using two diverse methods CPR (Constant Packet Re-arranging) and WCPR (Without Constant Packet Re-arranging) in diverse platforms. This paper proposes a new version of the TCP which gives the high throughput when the packet rearranging occurs and in another case if the packet rearranging is not occurs then in that case also it is friendly to other version of the TCP. The key feature of Constant packet rearranging is that duplicate ACKs are not used as an indication of packet loss. Instead the timer is used to detect the packet loss From a computational view-point, CPR is more demanding than WCPR. Because CPR does not rely on duplicate acknowledgments, packet rearranging (including out-or-order acknowledgments) has no effect on CPR performance.

Keywords

CPR, WCPR, Congestion Control, Packet Rearranging.

Full Text

   

Anurag Lal ¹

Affiliations
1 CSE Department, CIT, Rajanadgaon, IN

Abstract

There are different techniques that can be used to recognize handwritten digits and characters. Two techniques discussed in this paper are: Pattern Recognition and Artificial Neural Network. Both techniques are defined and different methods for each technique is also discussed. Bayesian Decision theory, Nearest Neighbor rule, and Linear Classification or Discrimination is types of methods for Pattern Recognition. Shape recognition, Character and Handwritten Digit recognition uses Neural Network to recognize them. Neural Network is used to train and identify written digits. After training and testing, the accuracy rate reached 99%.This accuracy rate is very high.

Keywords

Pattern Recognition, Multilayer, Cluster Neural Network.

Full Text

   

Anurag Lal ¹

Affiliations
1 CSE Department, CIT, Rajanadgaon, IN

Abstract

Network Intrusion Detection aims at distinguishing the behavior of the network. It is an inseparable part of the information security system. Due to rapid development of attack pattern it is necessary to develop a system which can upgrade itself as new threats are detected. Also detection rate should be high because the rate with which attack is carried out on the network is very high. In response to this problem AdaBoost Based Algorithm is proposed which has high detection rate as well as low false alarm rate. In this algorithm decision stumps are used as weak classifier. The decision rules are provided for both categorical and continuous features. Weak classifier for continuous features and weak classifier for categorical features are combined to form a strong classifier. Strategies for avoiding the over fitting are adopted to improve the performance of the algorithm.

Keywords

Intrusion Detection System, Adaboost Algorithm, Security, Machine Learning, Neural Networks.

Full Text