Informatics Publishing Limited

K. Vithiya Ruba ¹, D. Venkatesan ¹

Affiliations
1 School of Computing, SASTRA University, Tamil Nadu - 613 40, IN

Abstract

Twitter is a popular micro-blogging platform which allows the users to share their opinion on any domain. The thoughts of the people vary according to the domain and also the opinion may contain both positive and negative words which are called as opinion words and are given in the form of dictionary called lexicon dictionary. The sentiment analysis done without feature extraction fails to give the deep result about the users opinion but in our proposed approach , features of the domain are extracted by building ontology which helps in getting the refined sentiment analysis. Feature based sentiment analysis gives the best result. While analyzing the sentiment, scores are assigned to the tweets so that the sentiment score of our tweets are compared with the third party like American Customer Satisfaction Index score. This comparison shows that our score assignment gives the detailed analysis of the features than the third party.

Keywords

Opinion Words, Ontology, Protégé, Sentiment, Sentiment score, Tweets, Twitter

Full Text

   

S. Indhumathi ¹, D. Venkatesan ¹

Affiliations
1 School of Computing, SASTRA University, Thanjavur - 613401, Tamil Nadu, IN

Abstract

Wireless Sensor Network is a group of sensor nodes aimed to monitor various environmental conditions at the field locations. Eventually the sensed data is sent to a processing center. Each sensor node consists of a different number of sensors for sensing different parameters in the field area. The sensor nodes are usually equipped with a pair of limited alkaline battery. Achieving maximum coverage deployment is a major problem in wireless sensor networks. In the Existing system, by using the Random Deployment some nodes may get overlapping this will cause unbalanced structure. By using maximum coverage sensor deployment problem, the Coverage achieved only for static nodes. In the proposed system, a Genetic Algorithm is used to deploy Sensor Nodes for the Maximum coverage within the area, where the sensors are of different types. In this work, first analyze the total coverage area the WSN, identify the types of Sensor nodes and Coverage sensing distance, and calculate the coverage sensing distance for the combination of all sensor types based on radius of each node. Improving the deployment of Dynamic nodes for achieving maximum coverage deployment by using Genetic Algorithm. As a result, we were implemented this work in Java. This will show the best performance in coverage and network lifetime.

Keywords

Coverage Deployment, Genetic Algorithm (GA), Wireless Sensor Networks (WSN)

Full Text

   

P. Monika ¹, D. Venkatesan ¹

Affiliations
1 School of Computing, SASTRA University, Thanjavur - 613401, Tamil Nadu, IN

Abstract

Data mining is an emerging technology for applications such as text based mining, web based mining and it performs a major role in various domains for numerical data analysis, data statistics and machine learning. In this paper, data mining is used in machine learning of ANN (Artificial Neural Network). The Pruning technique in an MLP (The Multilayer Perceptron) neural network is to remove the unwanted neurons based upon their corresponding weights; as a result it improves the accuracy and speed of the network. In the existing system, based on their synaptic weights the pruning is performed and it removes the lowest weight neuron form the network. The result obtained from the existing method does not produce an optimized removal of the neuron. In the proposed system pruning is performed by using divisive clustering in MLP neural network. The main purpose of the Divisive algorithm in ANN is to split each neuron weight into sub neuron up to the fixed level and then remove the least weighted hidden neuron. The proposed method is implemented using the Java language. The Performance result obtained from the proposed method shows that it reduces the error rate and improves efficiency and accuracy of the MLP network. The present results confirm that DI-ANN (Divisive Artificial Neural Network) can provide a fast, accurate, and consistent methodology applicable to the neural network.

Keywords

DI-ANN (Divisive Artificial Neural Network) Algorithm, MLP (Multi Layer Perceptron), Neural Network (NN), Pruning Method

Full Text

   

S. Gayathri ¹, D. Venkatesan ¹

Affiliations
1 School of Computing, SASTRA University, Thanjavur - 613401, Tamil Nadu, IN

Abstract

Objectives: Watermarking technique on digital images with improved robustness and with optimum results can be achieved by using the intelligent technique Particle Swarm Optimization (PSO). Discrete Wavelet Transform (DWT) is used to uncover perceptually significant coefficients from the digital image. Methods: PSO is one amongst the efficient techniques in optimizing a problem having many numbers of candidates. This robust technique consists of two stages namely embedding and extraction stages. Insertion of watermark into the image is done at the embedding stage. Before embedding watermark into the image perform DWT and select any one of the bands LL, LH, HL and HH. Initially select the coefficients randomly and the optimal DWT coefficients for embedding the watermark at various positions are identified by PSO. Findings: In watermark extraction stage, the embedded watermark is identified by doing the reverse process. Robustness of the digital image watermarks implies the strength of the watermark against various image processing attacks and this can be checked by evaluating Cross Correlation value that is Normalized (NCC) and for fidelity Peak Signal to Noise Ratio (PSNR). It is evident that the robustness of the image is well maintained by this technique. As a further improvement any of the other intelligent technique namely Ant Colony Optimization can be used for optimization evaluation. Applications: Digital image watermarking plays a vital role in copyright infringement issues and it helps in predicting the authorized user by the digital signature that is watermarked in the image.

Keywords

Discrete Wavelet Transform, Fidelity, NCC, Particle Swarm Optimization, PSNR, Robustness, Watermarking.

Full Text

   

M. Rajeswari ¹, D. Venkatesan ¹

Affiliations
1 School of Computing, SASTRA University, Thirumalaisamudram, Thanjavur − 613401, Tamil Nadu, IN

Abstract

Objectives: User review is recognized as an effective way to promote a product or service. These reviews may be presents in different platforms such as blog, review site, social networking site, and messaging instance site. Methods: Selection of the best platform among different alternatives is called as Multi Criteria Decision Analysis/Method (MCDM) method. Objective is to select an effective platform for review. Fuzzy Analytic Hierarchy Process (AHP) is used for ranking and determines effective platform for reviews. It uses multiple criteria for making decision. In this research, the Fuzzy Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method is used which improves the quality of decision making for ranking alternative. Findings: When comparing with Fuzzy AHP method Fuzzy TOPSIS has an adequacy to include and exclude the criteria and alternative, which may change over a period of time. There is no restriction in number of criteria when we use the Fuzzy TOPSIS method. For future enhancements of this project can be done by using either fuzzy Analytic Network Process (ANP) or Aggregated Indices Randomization Method (AIRM) or can use hybrid method combination of any MCDM technique to rank the platform for product review. Applications: Fuzzy TOPSIS is useful for movie and product review and also we can apply this technique in uncertain situation.

Keywords

electronic Word Of Mouth (eWOM), Fuzzy Analytic Hierarchy Process (AHP), Fuzzy Scale, Fuzzy Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), Multi Criteria Decision Analysis/Method (MCDM), VIKOR.

Full Text

   

A. S. Suryapattatharasi , B. Amutha , D. Venkatesan

Abstract

Generally, the unicast and the multicast routing protocol are used to deliver messages between randomly chosen nodes. The major problem with these type of routing is that they do not process the distributed information's. Several algorithms or protocols such as shortest routing protocol and position based protocol are found to ensure effective routing in sparse network. These techniques focus on delivering the messages to certain nodes. However these techniques or protocols limit the process of retaining the information regarding the missing nodes. The proposed boomerang protocol effectively receives the messages from the nodes at any geographical area. The boomerang protocol attempts to utilize the physical location based information to increase the performance of geo cast delivery. A unique feature of this protocol is that it records the geographical trajectory while moving away from the origin, thus paving a way to exploits the recorded trajectory in case of returning back. This protocol simply attempts to provide more efficient routing method while maintaining the information of the missing node. Boomerang protocol finds a wide application in LAN, WAN, and other mobile communication network.

Keywords

Geo Cache, GPS, Mobile, Location Aware Information, Infrastructure Less Data Management

Full Text

   

J. Aravind Kumar ¹, D. Joshua Amarnath ¹, D. Venkatesan ¹, K. Vijai Anand ²

Affiliations
1 Department of Chemical Engineering, Sathyabama Institute of Science and Technology, Chennai - 600119, Tamil Nadu, IN
2 Department of Physics, Sathyabama Institute of Science and Technology, Chennai - 600119, Tamil Nadu, IN

Abstract

Polycyclic Aromatic Hydrocarbons (PAHs), group of structurally related chemicals consisting of aromatic rings with no substitutions form a potent class of environmental pollutants. Removal of PAHs is necessary because it causes carcinogenic and mutagenic effects. Adsorption and Fenton’s process is chosen among all the other methods. Among the various adsorbents, palm shell activated carbon gave the best adsorption. But comparatively the efficiency of Fenton’s process is better than adsorption. Optimization of parameters such as temperature, time and dosage, concentration, pH are reported.

Keywords

Anthracene, Adsorption, Fenton’s Process, Palm Shell.

Full Text