Computer Engineering

The area of sentiment mining (also called sentiment extraction, opinion mining, opinion extraction, sentiment analysis, etc.) has seen a large increase in academic interest in the last few years. Researchers in the areas of natural language processing, data mining, machine learning, and others have tested a variety of methods of automating the sentiment analysis process. The feasibility and the benefits of the proposed approaches are demonstrated by means of spam review that is widely used in the field of sentiment classification. In the proposed work, a comparative study of the effectiveness of ensemble technique is made for sentiment classification. Bagging and boosting are two relatively new but popular methods for producing ensembles. In this work, bagging is evaluated on spam review data set in conjunction with Naive Bayes (NB), Support Vector Machine (SVM), Genetic Algorithm (GA) as the base learners. The proposed bagged NB, SVM, GA is superior to individual approaches for spam review in terms of classification accuracy.

A distributed adaptive opportunistic routing scheme for multihop wireless ad hoc networks is proposed. The proposed scheme utilizes a reinforcement learning framework to opportunis-tically route the packets even in the absence of reliable knowledge about channel statistics and network model. This scheme is shown to be optimal with respect to an expected average per-packet re-ward criterion. The proposed routing scheme jointly addresses the issues of learning and routing in an opportunistic context, where the network structure is characterized by the transmission success probabilities. In particular, this learning framework leads to a sto-chastic routing scheme that optimally “explores” and “exploits” the opportunities in the network.

In the recent era, the extensive and rapid growth in internet technology is creating a important need to develop several newer techniques to protect authentication, ownership, copyright and content integrity of digital data. A proposed solution to this kind of problem is to digital watermarking. Digital watermarking has viable technique to the need of authentication, ownership, copyright and content integrity of digital data. Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT) are two most popular tools used in watermarking algorithm to protect digital image. Aim of this paper is to provide robust technique based on combined DWT and DCT propose a new watermarking algorithm for digital image. Discrete wavelet transform is used to provide robustness and discrete cosine transform has shorter elapsed time, taking advantages of both this technique proposed a new digital image watermarking algorithm. Results show that this algorithm combines the advantages of these two transforms.

Area Aided steering convention utilizes area data to diminish directing overhead of the specially appointed system regularly the LAR convention utilizes the GPS (Global Positioning System) to get these area information's. With the accessibility of GPS, the versatile hosts knows there physical area. In a versatile impromptu system comprises of remote has that may move frequently. Development of hosts brings about a change in courses, obliging some instrument for deciding new courses. A few steering conventions have as of recently been proposed for specially appointed systems. This paper proposes a methodology to use area data (case in point, acquired utilizing the worldwide situating framework) to enhance execution of directing conventions for impromptu systems. By utilizing area data, the proposed Location-Aided Routing (LAR) conventions confine the quest for another course to a more diminutive "appeal zone" of the impromptu system. This effects in a noteworthy lessening in the amount of steering messages. We exhibit two calculations to focus the solicitation zone, and likewise propose potential improvement to our calculations. One more test in outline of these systems is their powerlessness to security strike. We utilized a portion of the best cryptographic plans, for example, edge cryptography, to fabricate a very secure and profoundly accessible key administration, which structures the center of our security structure.

Sometimes BMI at almost the marginally end values of the standard prescribed intervals may wrongly categories a person either into normal or overweight category. This may be sometimes due to either wrong measurement of BMI value or wrong diagnosis of medical practitioner. This may lead to the wrong treatment of the patient. This is a frequent event everywhere in the world. To overcome such type of problems, it is observed that a soft computing approach can only provide a better satisfaction. Since it is a medical diagnosis problem, and since for almost all the medical problems the concerned physicians do always have some amount of hesitation (at best it could be nil), we have no other soft computing theory as our option but the intuitionistic fuzzy set theory of Atanassov only, for approaching towards a satisfactory solution.

Cloud computing is a paradigm from traditional computing, offers services over the Internet. The education system evolved in the culture, experiencing the paper maintained records for the report and all critical information of individual school management. In our research work, the system design for School management's information have been proposed and implemented through GCM methodology, where it has been molding as an application with the help of servers, to notify the student’s information like academic progress and enable to track them online by the parents, with accurate results. The scheme makes a more reliable, flexible approach to deliver up to date information to the ward's parents.

This Privacy preserving is most popular study for the research field. Privacy means gives the protection to the private information at preserving time. In Market place, discovery of frequent item sets using association rules mining is one of most important tasks in mining. Association rules is very helpful for finding frequent item sets to predicate about which item sets purchased together in a market and generate qualitative information that is useful for decision making. For distributed environment, database may be distributed as horizontally, vertically or mixed in computer network. The main problem in secure mining with the help of association rules is that transactions are distributed as vertically and the various sites want to find frequent item sets by participating themselves without discovering their individual data. The proposed method will find frequent item sets for vertical distributed database with the help of data miner using encryption based technique. Each sites prepare matrix with the local frequent item sets as per minimum support and encrypted it than send to other sites. The Scalar product with Boolean matrix is used for finding frequent item sets with secure computation between multiple sites without disclosing private input which improved efficiency and privacy of system.

Medical diagnosis is very much essential for the human beings to survive. Earlier the diagnosis, lesser is the pain. The disease unattended will lead to extreme conditions and greater pain, loss of health and high expenditure. This paper focuses on the comprehensive scalability analysis conducted on two kinds of thyroid dataset: original and synthetic. The results obtained with original dataset consisting of 215 instances with 5 features have been compared with the results obtained with synthetic dataset consisting of 1075 instances with 25 features. Initially the features were reduced to 3 features and 13 features on both the cases respectively. Subsequently classification accuracy is obtained using top most performing classifiers namely C4.5 and Multi layer perceptron. The results obtained are promising and the dimensionality reduction technique has proved better for smaller datasets as well as for larger datasets.

Environmental pollution has been a major irritant to industrial development. Chemical and chemical based industries are prime targets of the environmentalists for their crusade against pollution. Leather industry contributes to one of the major industrial pollution with the generation of pollutants causing chemicals namely lime, sodium sulphide, salts, organic solvents and phenolic compound. Leather industry has faced serious challenges on phenolic compounds. Thus this work focused on evaluation of biodegradation of phenol by using microaerophilic bacterial consortium to overcome the deleterious effect of chemical effluents to the environment. This biodegradation process will be ecofriendly and cost effective. This could be one of the ways of solving the industrial pollution resulting from tannery effluent. Tannery effluent with high phenol level fed in both batch reactor and continuous reactor by using sustainable Microaerophilic Bacterial consortia to degrade the phenolic compounds.

In the world wide network, there are several websites, in which each and every system has to be protected securely. The DNS Security Extensions (DNSSEC) make DNS the first core Internet system to be protected using public key cryptography. The DNSSEC not only protects the DNS, but has created interest in using this secured global database .There are several websites which are protected using different security schemes and are made less prone to the intruder .To save the data from misused, the admin control are also needs to be protected from viewing the public data. And to enforce these techniques, we need to fully encrypt the data and the database from viewing the actual data and transform into unusual data .Therefore, the data has been hided and protected from general view. In this paper, I propose to achieve a new theoretical model, called Public Data, which is robust and free from verification deployments as Communities of Trust (CoTs) by which it makes them secure . A reasonable DNSSEC deployment model and a typical choice of a CoT, has been properly implemented to protect the full data access for the general view of the user. Our limited deployment of Vantages has outperformed the verifiability of DNSSEC and has properly validated .Then the various adversary would be able to Man-in-middle attacks on arbitrary traffic into autonomous systems. Then the whole database and the backend process are encrypted and are in non-viewable formats.