In this paper, we propose a new approach of fault detection and diagnosis combining a Neural Nonlinear Principal.
Component Analysis (NNLPCA) and Partial Least Square (PLS). We have made a comparative study between the Linear Principal Component Analysis (LPCA) and Nonlinear Principal Component Analysis (NLPCA) to monitor a manufacturing process. This study has shown the capability of NLPCA in explaining nonlinear correlations in the process data. The traditional LPCA is limited to complex nonlinear systems; therefore, an adaptive NLPCA based on an improved training auto-associative neural network is presented. The proposed approach is applied to fault detection of a manufacturing process. The performance of the proposed approach is then illustrated and compared to those of classic LPCA.

A self-protection technique is suggested for adhoc network fall short of the objective of data privacy, data integrity, and authentication. Various security standards such as IEEE 802.11i, WPA, IEEE 802.1X were suggested to enhance the security issues in 802.11.Despite their efficiency, these standards does not provide any security approach for monitoring of these authentication in a distributed architecture. For the efficient monitoring of the authentication issue in adhoc network, in this paper we present a self monitored security approach for self-monitoring of key authentication for security protocol in adhoc networks. The processing overhead for the suggested approach is evaluated for a threshold based cryptographic approach.

Data mining is the process of extracting patterns from data. Data mining is becoming an increasingly important tool to transform this data into knowledgeable information. Clustering is a popular mining strategy that separates those data into subsets called clusters. The two main approaches of clustering is soft clustering and hard clustering. This project aims to implement soft clustering and hard clustering to enhance CRM (Customer Relationship Management). CRM is the process of managing a good relationship with customer and improve their profitability of their interactions with the customer. Data mining is an efficiently used tool in CRM. The telecom industry specifically handles the large data being generated by applying various data mining techniques. This project primarily focuses on applying apt clustering algorithm to identify the soft partitions of the customers namely the fuzzy c means (FCM) clustering algorithm to determine the churn ratio accurately. 
Keywords-Telecom; Data mining; Customer Relationship Management; Clustering; K-Means; Fuzzy C Means

In today’s scenario as the usage of ISM band is increasing rapidly there are many scenarios where we need communication systems like wireless local area networks (WLANs) based on IEEE 802.11b specifications and wireless personal area networks (WPANs) based on IEEE 802.15.4 specifications coexisting in the same place. In case of such coexistent heterogeneous networks we have performance degradation.  In this paper, we propose a new scheme using channel scheduling for enhancing the performance of the networks. To evaluate the effectiveness of the proposed scheme, the performance metrics such as throughput, average end-end delay and average jitter is measured using Qualnet 4.5 simulation software. The simulation results show that the proposed scheme is effective in performance improvement for coexistence network of IEEE 802.15.4 for various topologies. 

Denoising of images using hybrid approach has been a predominant scenario in the current trends. The hybrid approach refers in giving up a combined framework to utilize the advantage of both Fourier and wavelet domain schemes. The noisy images including their Region of Interest (ROI) can yield encouraging results, when obtained through the proposed hybrid approach when comparing to the single transform domain method. Thus the motivation for the hybrid method has been the realization that shrinkage in single transform cannot yield good estimates in deconvolution problems. The main focus of this paper is denoising of degraded images in which the blurring operator inversion is followed by noise attenuation through scalar shrinkage in both the Fourier and wavelet domain. The Fourier shrinkage utilizes the structure of the noise inherent in deconvolution. While the wavelet shrinkage utilizes the piecewise smooth structure of real-world signals and images. The advantage of denoising in Fourier domain is the ability to deal with overlapping responses. But in some exceptions when Fourier domain is alone used, noise amplification might be a disadvantage. Noise can be reduced in the frequency domain by shrinking the frequency coefficients, but noise and signal may be difficult to separate. Wavelet transforms have advantages over traditional Fourier transforms for representing functions that have discontinuities and sharp peaks, and for accurately reconstructing finite, non-stationary signals. Hence in this paper it is solved by using the combined Fourier and Wavelet shrinkage.
Keywords—Deconvolution, Fourier domain, Wavelets, Restoration, Smoothness, Estimation.

In this project we deal fully about the security which has become one of the major issues for data communication over wired and wireless networks. Different from the past work on the designs of cryptography algorithms and system infrastructures, we will propose a dynamic routing algorithm that could randomize delivery paths for data transmission. The algorithm is easy to implement and compatible with popular routing protocols, such as the routing Information Protocol in wired networks and Destination-Sequenced Distance Vector protocol in wireless networks, without introducing extra control messages. An analytic study on the proposed algorithm is presented, and a series of simulation experiments are conducted to verify the analytic results and to show the capability of the proposed algorithm. In the past decades, various security-enhanced measures have been proposed to improve the security of data transmission over public networks. Existing work on security enhanced data transmission includes the designs of cryptography algorithms and system infrastructures and security enhanced routing methods. The main objective of the project is to propose a dynamic routing algorithm to improve the security of data transmission.  

Mobile Ad hoc Networks (MANETs) are multi-hop wireless networks consisting of radio-equipped nodes that are mobile. The topology of a mobile ad hoc network changes frequently and this dynamic topology makes it difficult to design an efficient routing protocol. Number of nodes in a MANET can be varied from a few nodes to hundreds and even more nodes. Hence it is important for a MANET routing protocol to keep its performance in an acceptable level increasing number of nodes. Since nature is source of many well-established scalable mechanisms, this paper inspires from nature to design its scalable routing protocol. This protocol is based on TORA routing protocol and borrows some principles from bee and ant colonies. Simulation results signify that NISR is very competitive and outperforms TORA in terms of total delivered data, network life time and system life time.

Keywords— Ad Hoc Networks, Scalable, Routing, TORA, Bee colony, Ant colony.

Software development effort prediction is one of the most key activities in software industry. Many models have been proposed to build a relationship between software size and effort; however we still have problems for effort prediction. This is because project data, available in the primary stages of project is often inadequate, unpredictable, uncertain and unclear. The need for accurate effort estimation in software industry is an ongoing challenge. Artificial Neural Network models are more apt in such situations. The present paper is concerned with developing software effort prediction models based on artificial neural networks. The models are designed to improve the performance of the network that suits to the COCOMO Model. Artificial Neural Network models are created using Radial Basis and Generalized Regression. A case study based on the NASA 93 database compares the proposed neural network models with the Intermediate COCOMO. The results were analyzed using different criterions VAF, MMRE, MARE, VARE, Mean BRE and Prediction. It is observed that the Radial Basis Neural Network provided better results.

Keywords— Intermediate COCOMO, Cost Estimation, Radial Basis Neural Networks, Generalized Regression Neural Networks.