YMCA University of Science and Technology

About: YMCA University of Science and Technology is a education organization based out in Faridabad, India. It is known for research contribution in the topics: Web page & Web crawler. The organization has 299 authors who have published 568 publications receiving 4547 citations.

Semantic Similarity between Web Documents Using Ontology

Abstract: The World Wide Web is the source of information available in the structure of interlinked web pages. However, the procedure of extracting significant information with the assistance of search engine is incredibly critical. This is for the reason that web information is written mainly by using natural language, and further available to individual human. Several efforts have been made in semantic similarity computation between documents using words, concepts and concepts relationship but still the outcome available are not as per the user requirements. This paper proposes a novel technique for computation of semantic similarity between documents that not only takes concepts available in documents but also relationships that are available between the concepts. In our approach documents are being processed by making ontology of the documents using base ontology and a dictionary containing concepts records. Each such record is made up of the probable words which represents a given concept. Finally, document ontology’s are compared to find their semantic similarity by taking the relationships among concepts. Relevant concepts and relations between the concepts have been explored by capturing author and user intention. The proposed semantic analysis technique provides improved results as compared to the existing techniques.

SMC based shunt active harmonic filter for power quality enhancement

Abstract: The purpose of this paper is to design a high performance sliding mode controller through the use of a new switching function and Positive Sequence Voltage Detector (PSVD). The controller aims at controlling the source current harmonics, compensating reactive power, and improving power factor. In SMC used, the signum function is replaced by saturation function to smooth out the signal for minimizing chattering effect. By incorporating PSVD, the controller can be used, especially in the case of high voltage distortion or unbalanced input voltage signals. The proposed method is tested with balanced and unbalanced linear/non-linear load.

A Review of Video Shot Boundary Detection (SBD) technique

Abstract: The first step in effective video management system which provides the facility of content based video search, analysis and retrieval system is Video Shot Boundary Detection (SBD). A shot is a sequence of frames obtained from a single camera without interrupting it. A video contains a lot of shots that are accompanied with transitions from one shot to another. When shot transition occurs over a few number of frames, it is called Cut transition and if transition occurs over a relatively larger number of frames, it is called Gradual transition. The inspiration for the SBD scheme is from real time based problems. Several SBD techniques have been proposed over the years which yield satisfactory results for Cut transition but Gradual transition detection still have scope for improvement. In this paper, some of the recent benchmark SBD techniques are implemented and compared on the basis of Performance Metrics such as Precision, Recall, F1 score and Time Complexity.

Comparative analysis of traditional and modern controller for piezoelectric actuated nanopositioner

Abstract: Nanopositioning, Precise control and manipulation of devices and materials at nanoscale is the most important constraint of nanotechnology. Nanopositioning stage consisting of piezoelectric actuator has widespread use in applications requiring positioning with precision at nanoscale. In this paper, open loop characteristics of the non minimum phase nanopositioning system are investigated. To guarantee high precision positioning and improvement in the system characteristics demands different types of feedback controllers. The main aim of the controller is to design a closed loop system with good dynamic characteristics and to maintain the desired stability margins. PID controller is a generic closed loop controller widely used for industrial control applications. This paper presents design of PID controller using Ziegler Nichols tuning method. This paper analyzes time and frequency response of traditional PID controller. Non minimum phase system can be stabilized more effectively by modern controller such as pole placement controller, Linear Quadratic Regulator (LQR) and H infinity controllers. This paper develops a methodology for the design of full state feedback controller using pole placement and LQR control techniques. Step response and frequency responses under variety of conditions are plotted and analyzed to verify the effectiveness of the proposed controllers. Closed loop system is analyzed in both time and frequency domain for different dynamic characteristics such as rise time, settling time, maximum overshot and stability margins. A comparative assessment based upon the system’s response characteristics of traditional PID controller, pole placement controller and LQR controller is presented. Simulation results for the performance analysis are carried out on MATLAB and demonstrate effectiveness and adaptability of controller for precise control of piezoelectric actuated nanopositioning system.

Context ontology driven relevant search using data mining techniques (CODT)

Abstract: The size of the publicly indexable World Wide Web (WWW) has probably surpassed 14.3 billion documents and as yet growth shows no sign of leveling off. Search engines encounter the problem of ambiguity in words; therefore, search engines use ontology to find pages with words that are syntactically different but semantically similar. The knowledge provided by ontology is extremely useful in defining the structure and scope for mining web content. Context-ontology is a shared vocabulary to share context information in a pervasive computing domain and include machine-interpretable definitions of basic concepts in the domain and relations among them. This paper proposes an architecture for relevant searching of web documents using data mining techniques such as clustering and association rules. These techniques together with context and ontology extract potentially useful documents from the database. Also, an algorithm has been devised which shows the working in sequence of steps. Finally, the results are compared with the prevailing approaches and with the help of an example it has been seen that CODT is better in context of relevancy.