A Novel Approach of Relay Selection using AF & DF Relaying Techniques
Abstract: The paper looks at the performance comparison of different detection schemes and also proposes how to group users at the relay to ensure mutual benefit for the cooperating users. This research work proposes a framework which shows single and multiple relay selection using amplify & forward and decoded & forward relaying techniques under Rayleigh fading environment for cooperative communication.

Authors: Rambhau Gaikwad, Abhishek Verma

File Name: Abhishek_80200-19-102.pdf
 Download Abstract | Download Full Paper

Sentiment Analysis on Twitter Data using SVM Classifier
Abstract: Analyzing the large volumes of data generated in social networks on public opinion about different topics can result in valuable discoveries. These activities are expensive to perform manually, they require many human resources and time. Sentiment analysis systems and data mining algorithms have proved to be very useful in order to obtain a general perception of the topics of interest and the opinion on them. In this paper we propose to analyze a set of data using a sentiment classifier to label publications made by users of social networks in conjunction with clustering algorithms to be able to detect which are the topics on which opinions are expressed. We propose to use a base of 2000 reviews of films labeled as positive and negative and then train support vector machine (SVM) classifier of sentiments. We performed our experiments using one thousand tweets. Experimental evaluations show that our proposed technique is more efficient and has higher accuracy compared to previously proposed methods.

Authors: Kanak Jagwani, Ishi Raghuvanshi, Janhvi Sharma

File Name: Kanak_80200-19-104.pdf
 Download Abstract | Download Full Paper

Energy Management in Wireless Sensor Networks using M-SEP and LEACH
Abstract: Thanks to the great steps taken in recent years in technological development, and in particular microelectronics and wireless communication techniques, small, networked and inexpensive communicating sensors are increasingly being used in industrial applications and applications, observation of the environment. However, the use of wireless sensor networks in such applications has to face several limitations imposed by sensors such as processing capacity, small memory size, and energy. Or the limits imposed by the network itself, such as the narrow bandwidth, the network dynamics due to the topological variation of the network and the appropriate communication protocols adapted to this type of network. In this paper, we test the Modified-Stable Election Protocol (M-SEP) and Low-Energy Adaptive Clustering Hierarchy (LEACH) under a few distinctive situations holding high level heterogeneity to low level heterogeneity. The performance comparison metrics are; network lifetime, network throughput and the number of alive nodes.

Authors: Shreya Jain, Sandeep Veerwani

File Name: Shreya_80200-19-103.pdf
 Download Abstract | Download Full Paper

A Novel Mathematical Model for Image Retrieval
Abstract: The paper addresses the problem of access to databases based on images in indexes calculated from the content of the image itself, known as Content-Based Image Retrieval (CBIR). Perform a review of the state of the art in this topic. With the help of colour, texture and shape features different set of pattern stored in the database, and according to the query image, the similar image categories of images are extracted out. Colour Similarity Metric is measured using Chi-Square, while Texture Similarity Metric and Shape Similarity Metric is measured using Euclidian Distance Metric. Image recovered using the colour, shape and texture measure and is done with Euclidian distance. The accuracy achieved nearly 90% of accuracy with Corel dataset.

Authors: Anil Mishra, Tanmay Kasbe

File Name: Anil_80200-19-105.pdf
 Download Abstract | Download Full Paper

Lung Cancer Image Classification system using Random Forest Classifier
Abstract: Lung cancer is a real public health problem. Indeed, it is the leading cause of cancer mortality in the world, survival at 5 years is only 15% and this is largely due to late diagnosis and high metastatic power. Improving the management of this type of cancer, therefore, implies better knowledge of the processes of oncogenesis and tumor invasion. Most of the models for lung cancer classification based on lung cancer images are various types of classification model with binarization image pre-processing. This paper proposes a method based on a Random forest classifier for lung cancer image classification from the given database images. Feature extraction of the image is accomplished using Gabor Wavelet and GLCM (Grey Level Co-occurrence Matrix). Then the extracted features are classified by the Random forest classifier. This paper provides the confusion matrix with sensitivity, specificity, and accuracy for Gabor wavelet, GLCM and Hybrid (Gabor + GLCM) based approaches.

Authors: Sudeep Gujar, Laal Singh Chauhan, Nisha Kumawat

File Name: Sudeep_80200-19-106.pdf
 Download Abstract | Download Full Paper