Character recognition is a field of machine learning that has been under research for several decades. The particular success of neural networks in pattern recognition and therefore character recognition is laudable. Research has also long shown that a single hidden layer network has the capability to approximate any function; while, the problems associated with training deep networks therefore led to little attention given to it. Recently, the breakthrough in training deep networks through various pre-training schemes have led to the resurgence and massive interest in them, significantly outperforming shallow networks in several pattern recognition contests; moreover the more elaborate distributed representation of knowledge present in the different hidden layers concords with findings on the biological visual cortex. This research work reviews some of the most successful pre-training approaches to initializing deep networks such as stacked auto encoders, and deep belief networks based on achieved error rates. More importantly, this research also parallels investigating the performance of deep networks on some common problems associated with pattern recognition systems such as translational invariance, rotational invariance, scale mismatch, and noise. To achieve this, Yoruba vowel characters databases have been used in this research.

This paper presents parametric fault diagnosis in mixed-signal analog circuit using artificial neural networks. Single parametric faults are considered in this study. A benchmark R2R digital to analog converter circuit has been used as an example circuit for experimental validations. The input test pattern required for testing are reduced to optimum value using sensitivity analysis of the circuit under test. The effect of component tolerances has also been taken care of by performing the Monte-Carlo analysis. In this study parametric fault models are defined for the R2R network of the digital to analog converter. The input test patterns are applied to the circuit under test and the output responses are measured for each fault model covering all the Monte-Carlo runs. The classification of the parametric faults is done using artificial neural networks. The fault diagnosis system is developed in LabVIEW environment in the form of a virtual instrument. The artificial neural network is designed using MATLAB and finally embedded in the virtual instrument. The fault diagnosis is validated with simulated data and with the actual data acquired from the circuit hardware.

Existing knowledge systems incorporate knowledge retrieval techniques that represent knowledge as rules, facts or a hierarchical classification of objects. Knowledge representation techniques govern validity and precision of knowledge retrieved. There is a vital need to bring intelligence as part of knowledge retrieval techniques to improve existing knowledge systems. Researchers have been putting tremendous efforts to develop knowledge-based system that can support functionalities of the human brain. The intention of this paper is to provide a reference for further research into the field of knowledge representation to provide improved techniques for knowledge retrieval. This review paper attempts to provide a broad overview of early knowledge representation and retrieval techniques along with discussion on prime challenges and issues faced by those systems. Also, state-of-the-art technique is discussed to gather advantages and the constraints leading to further research work. Finally, an emerging knowledge system that deals with constraints of existing knowledge systems and incorporates intelligence at nodes, as well as links, is proposed.

Providing spontaneously personalized services to users, at anytime, anywhere and through any devices represent the main objective of pervasive computing. Smart home is an intelligent environment that can provide dozens or even hundreds of smart services. In this paper, we propose an approach to present spontaneously and continuously the most relevant services to the user in response to any significant change of his context. Our approach allows, firstly to assist proactively the user in the tasks of his/her daily life and secondly to help him/her to save energy in the smart home environment. The proposed approach is based on the use of context history information together with user profiling and machine learning techniques. Experimental results show that our approach can efficiently provide the most useful services to the user in a smart home environment.

After generalization of Shannon’s entropy measure by Renyi in 1961, many generalized versions of Shannon measure were proposed by different authors. Shannon measure can be obtained from these generalized measures asymptotically. A natural question arises in the parametric generalization of Shannon’s entropy measure. What is the role of the parameter(s) from application point of view? In the present communication, super additivity and fast scalability of generalized hyperbolic measure [Bhatia and Singh, 2013] of probabilistic entropy as compared to some classical measures of entropy has been shown. Application of a generalized hyperbolic measure of probabilistic entropy in certain situations has been discussed. Also, application of generalized hyperbolic measure of fuzzy entropy in multi attribute decision making have been presented where the parameter affects the preference order.

Mining frequent and infrequent itemsets from a given dataset is the most important field of data mining. When we mine frequent and infrequent itemsets simultaneously, infrequent itemsets become very important because there are many valued negative association rules in them. Mining frequent Itemset is highly expensive, if the minimum threshold is low, whereas mining infrequent itemsets is highly expensive, if the minimum threshold is high. When the dataset size is very large, both memory usage and computational cost of mining infrequent items is very expensive. In addition, single processor’s memory and CPU resources are not enough to handle very large datasets. Parallel and distributed computing are effective approaches to handle large datasets. In this paper we proposed a method based on Hadoop-MapReduce model, which can handle massive datasets in mining infrequent itemsets. Experiments are performed on 8 node cluster with a synthetic dataset. The performance study shows that the proposed method is efficient in handling very large datasets.

The current disjoint path Ad hoc On-Demand Multi-path Distance Vector (AOMDV) routing protocol does not have any energy-awareness guarantees. When AOMDV is used in wireless sensor networks (WSNs) energy is an important consideration. To enhance the AOMDV protocol an extra energy metric is added along with the hop count metric. This Energy aware or EA-AOMDV improves path selection using a trade-off between energy and hop count, thus giving more longevity to WSNs. EA-AOMDV is compared to the current AOMDV routing protocol to prove its worth in the context of WSNs. It is found that EA-AOMDV leads to better WSN energy-awareness in resource constrained WSNs.

The problem with metaheuristics, including Teaching-Learning-Based Optimization (TLBO) is that, it increases in the number of dimensions (D) leads to increase in the search space which increases the amount of time required to find an optimal solution (delay in convergence). Nowadays, multi-core systems are getting cheaper and more common. To solve the above large dimensionality problem, implementation of TLBO on a multi-core system using OpenMP API’s with C/C++ is proposed in this paper. The functionality of a multi-core system is exploited using OpenMP which maximizes the CPU (Central Processing Unit) utilization, which was not considered till now. The experimental results are compared with a sequential implementation of Simple TLBO (STLBO) with Parallel implementation of STLBO i.e. OpenMP TLBO, on the basis of total run time for standard benchmark problems by studying the effect of parameters, viz. population size, number of cores, dimension size, and problems of differing complexities. Linear speedup is observed by proposed OpenMP TLBO implementation over STLBO.