Volume 9, Issue 5, May 2017

 

 

Fixed Intervals Approach: A Variant of Generalization for Improving the Privacy of E-Health Records
 

Fixed Intervals Approach: A Variant of Generalization for Improving the Privacy of E-Health Records

Pages: 76-86 (11) | [Full Text] PDF (200 KB)
A Majeed
School of Information and Electronic Engineering, Korea Aerospace University, Deogyang-gu, Goyang-si, Gyeonggi-do, 412-791, South Korea

Abstract -
The adoption of advance technologies in healthcare sector has brought about many improvements in the industry, including better communication between healthcare providers, improved quality of treatment and reduced cost. For the most part, these improvements have come about due to collaboration between healthcare providers and the sharing of healthcare data. However, this introduces various security and privacy concerns pertaining to the data in question. Preserving the privacy of the patients while simultaneously sharing data that would facilitate medical research is absolutely essential, for it is not just an ethical requirement but is also dictated by regulations. In this paper, we propose a new scheme of data privacy for e-health records which differs from existing approaches in its ability to prevent from identity disclosure even faced with adversaries having pertinent background knowledge. The proposed scheme is based on transforming data into fixed intervals and then replacing original values with averages. As a result, the proposed scheme offers improved data privacy than the current techniques available for publishing privacy preserving data.
 
Index Terms - Privacy, Healthcare, Anonymization, Collaboration, Background-knowledge

Citation - A Majeed. "Fixed Intervals Approach: A Variant of Generalization for Improving the Privacy of E-Health Records." International Journal of Computer Engineering and Information Technology 9, no. 5 (2017): 76-86.

Enhancing the Performance of CNNs using Evolutionary Programming
 

Enhancing the Performance of CNNs using Evolutionary Programming

Pages: 87-96 (10) | [Full Text] PDF (592 KB)
AM Elhady, E Radwan, HM El-bakry, AA Elfetouh
Faculty of Computer Sciences and Information Systems, Mansoura University, Egypt
Deanship of Scientific Research, Umm Al-Qura University, KSA

Abstract -
So far supervised learning of a dynamical system (Cellular Neural Network) is derived from a mathematical method that may lose some factors, like disturbance and noise data. This paper points to a new technique based on Evolutionary programming Computation (Genetic Programming) in order to overcome this problem. Because of the short in the prior techniques in counting the implementation problem, the hardware implementation constraints are taken into account when evolving new learning rules. Genetic programming is developed to discover the optimum supervised rules parameters and the form of the optimistic rule structure. With these new rules, the error rate and the learning time are reduced. The overall performance of CNNs is enhanced. Simulation results confirm the theoretical considerations. Moreover, comparisons with related work are given.
 
Index Terms - Genetic Programming, Cellular Neural Networks, Parametric Learning Rules, Hardware Implementation Constraints, Multi-Objective Fitness Function

Citation - AM Elhady, E Radwan, HM El-bakry, AA Elfetouh. "Enhancing the Performance of CNNs using Evolutionary Programming." International Journal of Computer Engineering and Information Technology 9, no. 5 (2017): 87-96.

Mathematical Analysis of the Relative Speedup Dynamics of the Service Layered Utility Maximization Model as Applied to Dynamic Workflow Oriented Webservice Composition
 

Mathematical Analysis of the Relative Speedup Dynamics of the Service Layered Utility Maximization Model as Applied to Dynamic Workflow Oriented Webservice Composition

Pages: 97-107 (11) | [Full Text] PDF (929 KB)
A Mulongo, E Opiyo, E Abade, W Odongo
InformationUAP-Old Mutual Group Limited, Nairobi, Kenya
School of Computing and Informatics, University of Nairobi, Nairobi, Kenya

Abstract -
The Service Layered Utility Maximization model, SLUM is an emerging two phase layered mixed integer programming (MIP) for solving the dynamic webservice composition problem far more efficiently than the state of the art. Recent performance studies have demonstrated that: One, SLUM relative speedup, , with respect to the standard one phase MIP algorithms, dynamically grows larger as the number of service providers per workflow task, grows larger and vice versa. Two, scales exponentially in the number of sequential worklow tasks, . However, the effect of varying degrees of service phase transition from layer one to layer two on has not been adequately explored. Secondly, existing studies have emphasized the layering scheme in which the size of layer one and the size of layer two are equal. However, in reality, and could differ depending on application scenarios. This paper formulates mathematical models that quantify the dynamics of SLUM relative speedup for any SLUM layering scheme of the form taking into account the degree of service phase transition, . For a known layering scheme , we derive mathematical functions for the minimum possible and for the maximum possible at known values of , , and , where is total number of webservice quality of service parameters. Third, we show that given , there are valid SLUM layering schemes. From the set of possible schemes, mathematical functions that capture the lower and upper SLUM speedup limits are presented. We pioneer a graphical method of visualizing the speedup dynamics as function of . At design time, service system architects could use these models to anticipate the runtime performance efficiency benefits of a selected SLUM layering scheme.
 
Index Terms - Mathematical Analysis, Performance Dynamics, Relative Speedup, Service Layered Utility Maximization, SLUM, Service Composition

Citation - A Mulongo, E Opiyo, E Abade, W Odongo. "Mathematical Analysis of the Relative Speedup Dynamics of the Service Layered Utility Maximization Model as Applied to Dynamic Workflow Oriented Webservice Composition." International Journal of Computer Engineering and Information Technology 9, no. 5 (2017): 97-107.