Cloud computing is an optimistic technology that leverages the computing resources to offer globally better and more efficient services than the collection of individual use of internet resources. Due to the heterogeneous and high dynamic nature of resources, failure during resource allocation is a key risk in cloud. Such resource failures lead to delay in tasks execution and have adverse impacts in achieving quality of service (QoS). This paper proposes an effective and adaptive fault tolerant scheduling approach in an effort to facilitate error free task scheduling. The proposed method considers the most impactful parameters such as failure rate and current workload of the resources for optimal QoS. The suggested approach is validated using the CloudSim toolkit based on the commonly used metrics including the resource utilization, average execution time, makespan, throughput, and success rate. Empirical results prove that the suggested approach is more efficient than the benchmark techniques in terms of load balancing and fault tolerance.

Osteoporosis disease is caused by hormonal changes, vitamin D, and calcium deficiency. With current technologies, the identification of osteoporosis requires many tests with the support of medications. Bone mineral density is a typical measure implemented using a DEXA scan which can be very costly. Such high technology equipment is usually not accessible for remote people, and thus a low-cost screening system is very appealing. This article proposes an osteoporosis prediction system that effectively determines its possibility of occurrence based on essential factors such as smoking habits and calcium level so that the people at high risk can be referred to access the DEXA scanner. Our proposed system is implemented by an improved version of the artificial immune system, enabling care providers to take precautionary measures at the right time to avoid the early development of osteoporosis. The experiments demonstrated a promising result of 94% prediction accuracy that proved its usefulness in identifying people with potential osteoporosis in the future.

In this Internet era, with ever-increasing interactions among participants, the size of the data is increasing so rapidly such that the information available to us in the near future is going to be unpredictable. Modeling and visualizing such data are one of the challenging tasks in the data analytics field. Therefore, business intelligence is the way in which a company can use data to improve business and operational efficiency whereas data analytics involves improving ways of making intelligence out of that data before acting on it. Thus, the proposed work focuses on prevailing challenges in data analytics and its application on social media like Facebook, Twitter, blogs, e-commerce, e-service and so on. Among all of the possible interactions, e-commerce, e-education, and e-services have been identified as important domains for analytics techniques. So, it focuses on machine learning technique in improving practice and research in such e-X domains. Empirical analysis is done to show the performance of proposed system using real-time datasets.

Network congestion remains one of the main barriers to the continuing success of the internet and Web based services. In this background, proxy caching is one of the most successful solutions for civilizing the performance of Web since it reduce network traffic, Web server load and improves user perceived response time. Here, the most popular Web objects that are likely to be revisited in the near future are stored in the proxy server thereby it improves the Web response time and saves network bandwidth. The main component of Web caching is it cache replacement policy. It plays a key role in replacing existing objects when there is no room for new one especially when cache is full. Moreover, the conventional replacement policies are used in Web caching environments which provide poor network performance. These policies are suitable for memory caching since it involves fixed sized objects. But, Web caching which involves objects of varying size and hence there is a need for an efficient policy that works better in Web cache environment. Moreover, most of the existing Web caching policies have considered few factors and ignored the factors that have impact on the efficiency of Web proxy caching. Hence, it is decided to propose a novel policy for Web cache environment. The proposed policy includes size, cost, frequency, ageing, time of entry into the cache and popularity of Web objects in cache removal policy. It uses the Web usage mining as a technique to improve Web caching policy. Also, empirical analyses shows that proposed policy performs better than existing policies in terms of various performance metrics such as hit rate and byte hit rate.

No abstract