Credit Card Fraud Detection and False Alarms Reduction using Support Vector Machines
In day to day life credit cards are used for purchasing goods and services with the help of virtual card for online transaction or physical card for offline transaction. In a physical-card based purchase, the cardholder presents his card physically to a merchant for making a payment. To carry out fraudulent transactions in this kind of purchase; an attacker has to steal the credit card. To commit fraud in these types of purchases, a fraudster simply needs to know the card details. Most of the time, the genuine cardholder is not aware that someone else has seen or stolen his card information. The only way to detect this kind of fraud is to analyze the spending patterns on every card and to figure out any inconsistency with respect to the “usual” spending patterns. To commit fraud in these types of purchases, a fraudster simply needs to know the card details. Most of the time, the genuine cardholder is not aware that someone else has seen or stolen his card information. The only way to detect this kind of fraud is to analyze the spending patterns on every card and to figure out any inconsistency with respect to the “usual” spending patterns. Fraud detection based on the analysis of existing purchase data of cardholder is a promising way to reduce the rate of successful credit card frauds. As manually processing credit card transactions is a time-consuming and resource-demanding task, credit card issuers search for high-performing and efficient algorithms that automatically look for anomalies in the set of incoming transactions. Data mining is a well-known and often suitable solution to big data problems involving risk such as credit risk modelling, churn prediction and survival analysis. Nevertheless, fraud detection in general is an atypical prediction task which requires a tailored approach to address and predict future fraud. Though most of the fraud detection systems show good results in detecting fraudulent transactions, they also lead to the generation of too many false alarms. This assumes significance especially in the domain of credit card fraud detection where a credit card company needs to minimize its losses but, at the same time, does not wish the cardholder to feel restricted too often. In this work, we propose a novel credit card fraud detection system based on the integration support vector machines.
Published by: Mehak Kamboj, Shankey Gupta
Author: Mehak Kamboj
Paper ID: V2I4-1145
Paper Status: published
Published: July 12, 2016
A Hybrid Approach for Enhancing Security in RFID Networks
RFID (Radio-Frequency Identification) is a technology for automatic identification of things and people. Human beings are skillful at identifying things under many different challenge circumstances. A bleary-eyed person can quickly pick a cup out of coffee on a cluttered breakfast dining table each day, as an example. Computer sight, though, executes jobs which are such. RFID might be considered an easy method of explicitly objects that are labeling facilitate their “perception” by processing devices. An RFID device frequently only called an RFID label isa microchip that is small for wireless information transmission. It is generally mounted on an antenna in a package that resembles an adhesive sticker that is ordinary. The word “RFID” to denote any RF device whose function that is main identification of an object or person. Thisdefinition excludes simple products like retail stock tags, which simply indicate their particular presence and on/off condition during the standard end of the practical range. It also excludes products being transportable smart phones, which do a lot more than merely identify by themselves or their particular bearers. Numerous cryptographic models of security neglect to show crucial features of RFID systems. A straightforward design that is cryptographic as an example, catches the top-layer communication protocol between a tag and audience. In the reduced layers are anticollision protocols along with other RF that is basic notably enumerate the safety dilemmas present at multiple interaction layers in RFID methods. This work proposes a hybrid that is brand new and AES based Encryption mechanism for RFID program.
Published by: Bhawna Sharma, Dr. R.K. Chauhan
Author: Bhawna Sharma
Paper ID: V2I4-1144
Paper Status: published
Published: July 12, 2016
Performance Analysis of Multi-Hop Parallel Free-Space Optical Systems over Exponentiated Weibull Fading Channels Optimize by Particle Swarm Optimization
The performance of multihop parallel free- space optical (FSO) communication systems with decode-and-forward (DF) protocol over exponentiated Weibull (EW) fading channels has been investigated. The ABER and outage probability performance are analyzed under different turbulence conditions, receiver aperture sizes and structure parameters (R, C). The ABER and outage probability for FSO system is derived based on PSO. The ABER performance of the considered systems are investigated systematically combined with MC simulations. The comparison between EW fading model and PSO based EW fading channel demonstrates that performance of the both systems could be enhanced by large aperture sizes with the structure parameters R and C. With the particle swarm optimization (PSO) optimize path selection, fast average bit error rate (ABER) and outage probability reduction are achieved.
Published by: Babita, Dr. Manjit Singh Bhamrah,
Author: Babita
Paper ID: V2I4-1143
Paper Status: published
Published: July 9, 2016
Implementation of OLSR Protocol in MANET
Mobile ad hoc networks (MANETs) are autonomously self-organized networks without infrastructure support. In a mobile ad hoc network, nodes move arbitrarily; therefore the network may experience rapid and unpredictable topology changes. Because nodes in a MANET normally have limited transmission ranges, some nodes cannot communicate directly with each other. Hence, routing paths in mobile ad hoc networks potentially contain multiple hops, and every node in mobile ad hoc networks has the responsibility to act as a router. In this paper, we implement the OLSR Protocol in MANET to know how much data sent by the OLSR in bits/sec.
Published by: Rohit Katoch, Anuj Gupta
Author: Rohit Katoch
Paper ID: V2I4-1142
Paper Status: published
Published: July 7, 2016
Tumor Segmentation and Automated Training for Liver Cancer Isolation
Image segmentation is the process of subdividing the image to into its parts that are constituent and is considered one of the most difficult tasks in image processing. It plays a task that is a must any application and its particular success is based on the effective implementation of the segmentation technique. For numerous applications, segmentation reduces to locating an object in an image. This involves partitioning the image into two classes, background or object. Into the individual system that is visual segmentation happens obviously. Our company is experts on detecting patterns, lines, edges and forms, and making decisions based upon the information that is visual. At that time that is same we have been overwhelmed by the quantity of image information which can be captured by technology, as it is not feasible to manually process all such images.Automatic segmentation of tumor faction from medical pictures is difficult due to size, shape, place and presence of other objects with the intensity that is exact same in the image. Therefore, cancer segments from the liver where tumor persists cannot be easily segmented accurately from medical scans utilizing approaches that are traditional. The performance of ANN been examined in classifying the Liver Tumor in this research. An approach for segmentation of tumor and liver from medical pictures is principally used for computer aided diagnosis of liver is required. The method is use contour detection with optimized threshold algorithm. The liver is segmented region that is utilizing technique efficiently close around the liver tumors. The whole process is a learning that is supervised; the classifiers require training information set which can be segmented. The classifier that is last evaluated with test set total error in tumor segmentation of this liver is be calculated. Algorithm should be based on segmentation of abnormal regions in the liver. The category regarding the regions can be carried out based on shape categorization and lots of other features using methods such artificial networks that are neural.
Published by: Shikha Mandhan, Kiran Jain
Author: Shikha Mandhan
Paper ID: V2I4-1141
Paper Status: published
Published: July 5, 2016
MRI Fuzzy Segmentation of Brain Tumor with Fuzzy Level Set Optimization
Image segmentation is a task that is fundamental many image processing and computer vision applications. Due to the existence of noise, low contrast, and intensity in homogeneity, it really is still a difficult issue in majority of applications. One of the steps that are first way of understanding images is to segment them in order to find down different objects inside them. However, in real images such as MRI graphics, noise is corrupting the image information or image usually consists of textured sections. The images produced by MRI scans are frequently grey images with strength in the product range scale that is gray. The MRI image associated with the brain comprises of the cortex that lines the surface that is outside of brain additionally the gray nuclei deep inside of the mind including the thalami and basal ganglia. As Cancer may be the leading cause of death for all as the explanation for the condition remains unknown, very early detection and diagnosis is one of the keys to cancer control, and it will increase the success of treatment, save lives and reduce expenses. Health imaging is very often used tools which can be diagnostic detect and classify defects. To eliminate the dependence of the operator and increase the precision of diagnosis system aided diagnosis computer are a valuable and ensures that are advantageous the detection of cancer tumors and classification. Segmentation techniques based on gray level techniques such as for instance threshold and methods based on region are the easiest and find application that is restricted. However, their performance can be improved by incorporating them with the ways of hybrid clustering. practices based on textural characteristics atlas that is using look-up table can have very good results on the segmentation of medical pictures , however, they require expertise within the construction of the atlas Limiting the technical atlas based is that , in some circumstances , it becomes difficult to choose correctly and label information has difficulty in segmenting complex structure with variable form, size and properties such circumstances it is best to use unsupervised methods such as fuzzy algorithms. In this work we proposed a novel fuzzy based MRI Image Segmentation algorithm, Fuzzy Segmentation involves the task of dividing data points into homogeneous classes or clusters making sure that things within the same class are as similar as possible and items in numerous classes are as dissimilar as you can.
Published by: Poonam Khokher, Kiran Jain
Author: Poonam Khokher
Paper ID: V2I4-1140
Paper Status: published
Published: July 5, 2016
