R. Rajkumar ¹, A. James Albert ², D. Chandrakala ³

Affiliations
1 Department of Mathematics, Kumaraguru College of Technology, Coimbatore- 641049, IN
2 Department of Science and Humanities, Hindustan College of Engineering and Technology, Coimbatore-641032, IN
3 Department of Computer Science, Kumaraguru College of Technology, Coimbatore-641049, IN

Abstract

Background: Weather forecasting has become one of the active research areas due to major scientific and technological problems. An analysis is made to utilize of data mining approaches in weather forecast. Methods: Fuzzy Neural Network (FNN) and Hierarchical Particle Swarm Optimization (HPSO) are used for predicting the weather changes occur in the atmosphere. The FNN uses biological neurons for exact calculation and the neural network adjust their weights by the practice of training. The HPSO is used here for better optimization of weights, also that to improve the overall performance Adaptive Inertia Weight Algorithm (AWA) is proposed. Results: As a result, the total error is reduced with little tolerance which results in accurate weather prediction. The network is trained in such a manner that the model will provide 94% of optimum results. Conclusion: The performances of the proposed algorithm were compared with other existing algorithm using other standard performance metrics which gives best results for the mean weather variables. The results show that the proposed approach is efficient in determining weather forecasting and moreover helps in climate change investigations.

Keywords

Fuzzy Neural Network, Hierarchy Particle Swarm Optimization Technique, Network Forecasting, Numerical Weather Prediction, Weather Forecasting

Full Text

   

Sathya Sankaran ¹, R. Rajkumar ¹

Affiliations
1 School of Computing Science and Engineering, VIT University, Vellore – 632014, Tamil Nadu, IN

Abstract

Objectives: In this work, a unique barcode for each patient reporting to a hospital is derived using the address details of the patient’s residence. Methods: The method adapted here is for identifying unique barcode detection error detection is done. A binary format of address is used for identification of patient’s records. Findings: It is useful for identifying patients and healthcare professionals inside the hospital. We propose a geographical calender and time barcodes which may be adopted universally.

Keywords

Barcode, Cloud, Error Detection, Primary Key.

Full Text

   

R. Rajkumar ¹, S. Saira Banu ¹

Affiliations
1 Karpagam University, Coimbatore - 641021, Tamil Nadu, IN

Abstract

Objective: The Impulse or salt and pepper noise corrupt the image data and quality of the image. There are several strategies in the existing algorithms to eliminate the noise but also have drawbacks. We identify the drawbacks of the standard algorithms in consider with apply in the proposed algorithm to achieve the prefect noise removal. Enhance the Leading Diagonal Algorithm with different approaches to improve and identify the betterment of the proposed algorithm and prove with the parameter analysis of standard filters. Materials and Methods: Images are corrupted by the high density level of impulse noise. Retrieving the images by the proposed algorithm to make the images in better quality and preserving the edges. The Proposed algorithm based on the 5th pixel is called the pixel of processing and consider the diagonal matrix elements of the finding window to effectively to eliminate the high density noise. The Lena, The baboon and the lady-image are taken as the sample images to mix with the percentage of noises. The restoration of the sample images with the comparison of standard or recommended filtering algorithm. With the help of PSNR and MSE parameter to prove the improved algorithm is better than the available filtering algorithms. Findings: The noise density from 10% to 50% is added to the sample images in the increment of every 10% of noise density. Based on the parameter analysis (PSNR and MSE) with the standard filters to dedicate the new algorithm is for the betterment of more than the available filtering algorithm. The new algorithm is improving the quality and preserving the edges. Conclusion: To get the higher PSNR value and reducing value of MSE to prove the new approach is for the enhancement of the quality and retrieval of the images from the high density corrupted noise.

Keywords

Diagonal Sorting Method, Impulse Detection, Median Filter, Mean Square Error, Peak Signalto Noise Ratio.

Full Text

   

Vasanthi Subramanian ¹, R. Rajkumar ¹

Affiliations
1 School of Computer Science and Engineering (SCOPE), VIT University, Vellore - 632014, Tamil Nadu, IN

Abstract

Background: Hindi being India's official language, it is critical to know how to improve teaching methods and strategies to effectively help non-native speakers to overcome the language barrier. Methods: The major purpose of the article is to investigate adaptive content m-learning system for learning and teaching basic Hindi. It is expected from the literature that the interactive teaching characteristics are activities of student-centered for communication during real-life sociable situations. Findings: Hopefully, this basic Hindi language learning system will be able to help the students learn the unknown language by relying on the system's interaction to augment their language learning. It gives teachers a practical knowledge to use the system and receive extra information. Applications: It helps to assess learners from different viewpoints and provides an in-depth understanding of students in different learning styles.

Keywords

Adaptive Content, Context Based, Hindi Language, M-Learning, Personalized Learning

Full Text

   

R. Nirmala ¹, R. Rajkumar ²

Affiliations
1 Department of Civil Engineering, Sathyabama University, Rajiv Gandhi Road, Jeppiaar Nagar, Chennai - 600119, Tamil Nadu, IN
2 Department of Civil Engineering, SSN College of Engineering, Chennai – 603110, Tamil Nadu, IN

Abstract

Background/Objectives: Understanding the mechanics of buried pipes is complicated due to the imprecision in the properties of the soil envelope. This paper deals with the study of behavior of flexible Un-Plasticized Poly Vinyl Chloride pipes buried in loose and dense sand backfill. Methods/Statistical Analysis: A design methodology is proposed for prediction of the performance of pipes using the finite element method ANSYS. The pipe soil interaction under static loading conditions is investigated. Height of the soil above the pipe varies with the ratios of diameter of the pipe. The numerical results are compared with the available values calculated using the theoretical approach Spangler Deflection theory. Findings: Provision of geogrid reinforcement above the crown of buried pipe is suggested to minimize the vertical deflection of buried Un-Plasticized Poly Vinyl Chloride pipes.

Keywords

Backfill, Pipe - Soil Interaction, Soil Envelope, Static Loading

Full Text

   

K. N. Asha ¹, R. Rajkumar ¹

Affiliations
1 School of Computing Science and Engineering, Vellore Institute of Technology, Vellore – 632014, Tamil Nadu, IN

Abstract

Objectives: In recent years, we have tremendous growth of users in Online Social Networks (OSN) such as Facebook, Google+, twitter etc. This becomes major reason for enabling web as largest market defining E-Commerce. Many companies use OSN as their sales channel as they can reduce operating cost for managing orders significantly compared with traditional channels. Also viral marketing is very popular in OSN. But the major drawback of this channel is if the company doesn’t satisfy the customer, then the same OSN can rapidly propagate a bad reputation of the company affecting the company’s business. Hence for customers it has become very important to identify and filter dishonest recommenders. So it becomes very important to recommend right items to right customers. In this survey, we aim to give a comprehensive review of research related to E-Commerce in OSN. Methods: First, we discuss web usage mining techniques for better shopping websites to satisfy customers. Also we discuss web mining techniques to find dishonest recommenders in OSN. Findings/ Improvements: Our survey explores the existing research highlights and also presents various researches in these topics. Also, we propose recommendation system which uses Semantic Web Mining process integrated with domain ontology which can be used to extract interesting patterns from, complex and heterogeneous data.

Keywords

E-Commerce, OSN, Recommender System, Spammer Detection, Web Mining Technologies

Full Text

   

N. Vivekananthamoorthy ¹, E. R. Naganathan ², R. Rajkumar ³

Affiliations
1 Information Technology, SRM University, Kattankulathur, Chennai - 603203, Tamil Nadu, IN
2 Computer Science and Engineering, Hindustan University, Padur, Chennai - 602203, Tamil Nadu, IN
3 Computer Science and Engineering, SRM University, Kattankulathur, Chennai -603203, Tamil Nadu, IN

Abstract

Background/Objectives: In recent times, there is phenomenal increase in usage of Social Networking Sites like Facebook, LinkedIn etc. by college students and young professionals. This study focuses on identifying key factors that influence LinkedIn usage and the role of eWOM communication in enhancing social connectivity and engagement of students in meaningful activities to improve their social and academic standings. A theoretical model on social networking by students is proposed and the results and recommendations of this study will be brought to practical use towards student empowerment. Methods/Statistical Analysis: A preliminary survey was conducted to understand how young university students use the Social Networking Site LinkedIn and the responses were used to frame a questionnaire. A second level survey was conducted among the same set of participants by collecting their responses in five point Likert Scale. Exploratory Factor Analysis was conducted using the LinkedIn Survey responses to identify the hidden factors associated with the indicator items in the data set. Subsequently, a theoretical model was constructed using Structural Equation Modeling principles, depicting the interrelationships between the latent constructs and indicator items constituting a measurement model and a structural model. Four Hypotheses were framed such that Social Media Usage and eWOM communication have significant positive effect on Student Empowerment. Finally, Confirmatory factor analysis was done to prove the hypotheses and to analyze how well the model fits into the theory. The software IBM SPSS, and AMOS 23 were used to perform multivariate statistical analysis on the LinkedIn Survey response items. Findings: The exploratory study on LinkedIn Usage Survey responses revealed three latent factors that accounted for 69.462 percent of the total variance. The three key factors explaining the eWOM behavior of students in LinkedIn usage were Expert Opinion Seeking, Networking with Professionals and Notification of Profile Changes. The latent factors and associated relationships were used to frame a theoretical model based on SEM techniques. Based on Confirmatory factor analysis done on this model using the data set revealed that the model supported the hypotheses H1, H2, H3 and H4 and all indicators in the model significantly loaded to their respective factors and the predicting variables had a significant positive effect on the predicted variable. The factor loadings were fair to excellent ranging from .634 to .853 and the test for model fitness showed good fitness result based on value of various fitness indices which were within accepted limits. Based on CFA, the important fitness indices and their values arrived at were: CMIN/df = 2.022, NFI = 0.824, TLI = 0.887, RMSEA = 0.098 and CFI = 0.901. Improvements/Applications: The accuracy of the predicting ability of the proposed theoretical model can be improved by augmenting this research study and statistical analysis to be extended to a larger target group belonging to different institutions to achieve good model fit as well as for testing the scalability of the model. As a future work, this model can be integrated with online learning systems also with the aim of improving student engagement in the current online learning scenarios.

Keywords

Confirmatory Factor Analysis, eWOM Communication, Exploratory factor Analysis, Multivariate Analysis, Online Social Networking Sites (OSN), Principal Component Analysis, Social Networking Sites, Structural Equation Modeling (SEM), Student Empowerment, Uses and Gratification Theory (U&G)

Full Text