Search

This paper identifies the supervised and unsupervised learning algorithms for imbalanced classes to extract rules for them. Most Machine learning algorithms work best when the number of instances of each class is roughly equal, but there are only specific algorithms to deal with the imbalanced classes. Imbalanced classification problems mean that the dependent (response) variable has an imbalanced proportion of classes. This results in biased predictions, misleading inaccuracies, and fluctuating performances in the datasets and henceforth, imbalanced classes need attention in machine learning. In the field of machine learning, the traditional model evaluation methods do not accurately measure the model performances when faced with imbalanced classes. However, in an imbalanced dataset, the minor class does not significantly contribute towards accuracy. Hence, accuracy should not be used to evaluate the models' performance for an imbalanced dataset. Thus, this paper discusses several data mining methodologies like Sampling, Clustering, Distance measures and Performance measures to extract more rules for the imbalanced classes.

Security and privacy remain a major challenge for data communication between an Air Controller Tower (ACT) and Flight mainly due to the immense scale of data transfer and distributed nature of this Internet of Things (IoT) network. Blockchain and its related approaches provide decentralized security and privacy, however yet they involve significant energy, delay, and computational overhead. The aforementioned costs are not feasible for most resource-constrained IoT devices. The proposed security models of this paper acknowledged these drawbacks by circumventing the computational resource-intensive Proof of Work (PoW) concepts. The models consist of three main components: Customized Digital Certificate (CC), Validator (Smart Contract), and Storage (Blockchain Network). A set of new algorithms are proposed to outline various phases during communication and embedding message security. The proposed models are evaluated in a simulated environment for secure data communication between Flights and ACTs. Each ACT and Flight need to be equipped with IoT devices and Validator (SC) that can send and receive validated messages and stored in the Blockchain network at both ends. This paper explores the communication realm and outlines Cryptography and Blockchain-based approaches. Extensive empirical work has been done to demonstrate that the proposed models are viable with respect to confidentiality, integrity, and availability. The presented models can be further utilized for tracing of message exchange for integrity and non-repudiation.

In the era of Big Data and Internet of Things, devices ubiquitously sense and gather data at a rapid pace. Various methods have been proposed to speed up the analysis of the data and also mining it for information. However, considering the resource constraints in majority of Internet of Things, this is challenging. The analysis of data can be handled using a massive array of compute nodes; but the data from the devices need to be transferred to the servers to start the process. This involves transmitting the data over the network. Of course, with the huge quantity of data, this requires significant energy and resources to do so. Therefore, in order to address such issues with data analytics using Internet of Things, using data compression techniques is a viable option. Since graphs represent most real world data, including data gathered by Internet of Things, methods to compress graphs have been in the forefront of such endeavors. In this paper we propose techniques to compress graphs by finding specific patterns and replacing those with identifiers that are of variable length, an idea inspired by Huffman Coding. Specifically, given a graph G = (V, E), where V is the set of vertices and E is the set of edges, and |V| = n, we propose methods to reduce the space requirements of the graph by compressing the adjacency matrix representation of the same. The proposed methods show up to 80% reduction is the space required to store the graphs as compared to using the adjacency matrix. The methods can also be applied to other representations as well. The proposed techniques help solve the issues related to transfer of data over the network in case of resource limited Internet of Things, and addresses the challenges of data analytics in this domain.

Detecting precise location of aircraft during the entire flight duration is a challenge in the domain of commercial aviation. Using radar and other available technology, flights operating entirely over land can be tracked easily. However, with long haul intercontinental flights, where majority of the flight path is over water bodies and out of range of radar, detecting the location of aircraft at all times is a challenge. In recent times, there have been disasters in commercial aviation, where an aircraft has gone missing. This has a huge social and financial impact on the specific airline and commercial aviation in general. Therefore, in this paper we study the problem of location detection for commercial aircraft and methods to improve location detection over any terrain the flight path traverses. We propose techniques based on the Internet-of-things (IoT) model for aircraft, where the aircraft can communicate with each other within a certain range. We introduce distributed algorithms to detect location using such methods that work effectively when the aircraft is outside the range of radar and on an oceanic route. Our results show that using the proposed methods, the precise location of all aircraft, including those intercontinental flights, can be tracked to a higher degree. Techniques to minimize the communication overhead introduced due to the proposed methods are also provided.

With the growth of commercial aviation over the last few decades there have been many applications designed to improve the efficiency of flight operations as well as safety and security. A number of these applications are based on the gathered data from flights; the data is usually acquired from the various sensors available on the aircraft. There are numerous sensors among the electrical and electronics devices on an aircraft, most of which are essential for the proper functioning of the same. With the sensors being operational throughout the time of movement of the aircraft, a large amount of data is collected during each flight. Normally, most of the gathered data are stored on a storage device on the aircraft, and are analyzed and studied later off-site for research purposes focusing on improving airline operation and efficiently maintaining the same. In certain cases, when there is data transfer during the flight, it is between the aircraft and an air-traffic-control (ATC) tower, which serves as the base station. The aircraft equipped with all these sensors, which can gather and exchange data, form a framework of Internet of things (IoT). Detecting and avoiding any form of turbulence for an aircraft is vital; it adds to the safety of both passengers and aircraft while reducing the operating cost of the airline. Therefore, in this paper, we study techniques to detect and avoid Clear Air Turbulence (CAT), which is a specific type of turbulence, based on the IoT framework of aircraft. We propose algorithms that consider both direct and indirect communication between aircraft within a specific region. Using simulation results, we show that our proposed techniques of direct communication using the IoT framework is faster than conventional techniques involving radio communication via both single ATC tower and multiple ATC towers.

Detecting the location of aircraft at all times during flight is of utmost importance in commercial aviation. With the increasing number of aircraft that operate over various routes across the globe, tracking each flight with accurate location detection is a challenge. In this paper, we propose an internet of things based framework for aircraft that can assist with tracking the same. Our model includes aircraft, air traffic control towers, project loon balloons and infrastructure enabled aircraft as the different components. Our proposed algorithms work for aircraft that operate on routes over land as well as large water bodies, specifically oceans. All the algorithms are implemented and tested using simulation for the tracking of aircraft in different scenarios. The experimental results show that using the proposed techniques, an additional 70% of aircraft under consideration can share location data and can be detected as compared to using conventional radar based techniques.

Techniques for predicting the efficiency of multi-core processing associated with a set of tasks with varied CPU and main memory requirements are introduced. Prediction of CPU and memory availability is important in the context of making process assignment, load balancing, and scheduling decisions in distributed systems. Given a set of tasks each with varied CPU and main memory requirements, and a multi-core system (which generally has fewer cores than the number of tasks), we provide upper- and lower-bound models (formulas) for the efficiency with which the tasks are executed. In addition, a model for average CPU availability is introduced from the empirical study for applications that require a single predicted value instead of bounds. To facilitate scientific and controlled empirical evaluation, real-world benchmark programs with dynamic behaviour (CPU and memory requirements change in a short interval of time) are employed on UNIX systems that are parameterised by their CPU usage factor and memory requirement.

Using Graphics Processing Units (GPUs) to solve general purpose problems has received significant attention both in academia and industry. Harnessing the power of these devices however requires knowledge of the underlying architecture and the programming model. In this paper, we develop analytical models to predict the performance of GPUs for computationally intensive tasks. Our models are based on varying the relevant parameters - including total number of threads, number of blocks, and number of streaming multi-processors - and predicting the performance of a program for a specified instance of these parameters. The approach can be used in the context of heterogeneous environments where distinct types of GPU devices with different hardware configurations are employed.

Digital image processing has become ubiquitous in our daily life and the demands to produce and process images are ever increasing. Large amounts of space are required to store these images. Image compression techniques are in high demand as they allow reduction in this storage space. The basis for image compression, as is for most other compression techniques, is to remove redundant and unimportant data. Lossless image compression techniques retain the original information in compact form and do not introduce any errors when decompressed. In this paper, we discuss such a lossless technique using a data structure that we name ldquogeneric Peano pattern mask treerdquo. It is an improvement over a previously discussed Lossless Image compression technique - ldquoPeano pattern mask treerdquo. Both these structures are based on the data structure - Peano mask tree.

dasiaSeven segment displaypsila is a very popular device that displays Indo-Arabic numerals by turning on or off the individual segments in it. Currently this display device is used to represent the numerals of almost all European languages. However Indian languages are lacking of a proper segmented display device to display its numerals. Over the past few years many research works have been done with Bangla language which result many proposals to display Bangla numerals. However we found very few significant works in displaying the numerals of other Indian languages. Also in the proposed devices for displaying Bangla numerals we found the number of segments are 16, 11, 10, 9, 8, 14 (dual 7 segments) etc. In this paper we are proposing a unified seven segment display device which is capable of displaying Bangla, English and most of the Indian languagespsila numerals that include Assamese, Manipuri, Hindi, Nepali, Marathi, Gujarati, Punjabi, Kannada, Telugu, Sikkim, Bhutan etc. Our proposed display device has the same simplicity that of English with an acceptable level of accuracy and uniformity.