Encoding Gene Expression Using Deep

Autoencoders for Expression Inference

Raju Bhukya

Department of Computer Science and Engineering, National Institute of Technology, India

Abstract: Gene expression of an organism contains all the information that characterises its observable traits. Researchers have invested abundant time and money to quantitatively measure the expressions in laboratories. On account of such techniques being too expensive to be widely used, the correlation between expressions of certain genes was exploited to develop statistical solutions. Pioneered by the National Institutes of Health Library of Integrated Network-Based Cellular Signature (NIH LINCS) program, expression inference techniques has many improvements over the years. The Deep Learning for Gene expression (D-GEX) project by University of California, Irvine approached the problem from a machine learning perspective, leading to the development of a multi-layer feedforward neural network to infer target gene expressions from clinically measured landmark expressions. Still, the huge number of genes to be inferred from a limited set of known expressions vexed the researchers. Ignoring possible correlation between target genes, they partitioned the target genes randomly and built separate networks to infer their expressions. This paper proposes that the dimensionality of the target set can be virtually reduced using deep autoencoders. Feedforward networks will be used to predict the coded representation of target expressions. In spite of the reconstruction error of the autoencoder, overall prediction error on the microarray based Gene Expression Omnibus (GEO) dataset was reduced by 6.6%, compared to D-GEX. An improvement of 16.64% was obtained on cross platform normalized data obtained by combining the GEO dataset and an RNA-Seq based 1000G dataset.

Keywords: Deep autoencoder, gene expression, internal covariance shift, machine learning, MLP, PCA.

Received March 5, 2019; accepted April 13, 2020

A study on Two-Stage Mixed Attribute Data Clustering Based on Density Peaks

Shihua Liu, Hao Zhang, and Xianghua Liu

Department of Information Technology, Wenzhou Polytechnic, China

Abstract: A Two-stage clustering framework and a clustering algorithm for mixed attribute data based on density peaks and Goodall distance are proposed. Firstly, the subset of numerical attributes of the dataset is clustered, and then the result is mapped into one-dimensional categorical attribute and added to the subset of categorical attribute data. Finally, the new dataset is clustered by the density peaks clustering algorithm to obtain the final result. Experiments on three commonly used UCI datasets show that this algorithm can effectively realize mixed attribute clustering and produce better clustering results than the traditional K-prototypes algorithm do. The clustering accuracy on the Acute, Heart and Credit datasets are 17%, 24%, and 21% higher on average than that of the K-prototypes, respectively.

Keywords: Mixed data clustering, density peaks, k-prototypes algorithm, validity index.

Received July 4, 2019; accepted September 27, 2020

Adaptive Optimization for Optimal Mobile Sink Placement in Wireless Sensor Networks

Arikrishnaperumal Ramaswamy Aravind¹ and Rekha Chakravarthi²

¹Department of Electronics and Communication Engineering, Sathyabama Institute of Science and Technology, India

²School of Electrical and Electronics, Sathyabama Institute of Science and Technology, India

Abstract: In recent years, Wireless Sensor Networks (WSN) with mobile sinks has attracted much attention as the mobile sink roams over the sensing field and collects sensing data from sensor nodes. Mobile sinks are mounted on moving objects, such as people, vehicles, robots, and so on. However, optimal placement of the sink for the effective management of the WSN is the major challenge. Hence, an adaptive Fractional Rider Optimization Algorithm (adaptive-FROA) is developed for the optimal placement of mobile sink in WSN environment for effective routing. The adaptive FROA, which is the integration of the adaptive concept in the FROA, operates based on the fitness measure based on distance, delay, and energy measure of the nodes in the network. The main objective of the research work is to compute the energy and distance. The proposed method is analyzed based on the metrics, such as energy, throughput, distance, and lifetime of the network. The simulation results reveal that the proposed method acquired a minimal distance of 24.87m, maximal network energy of 94.54 J, maximal alive nodes of 77, maximal throughput of 94.42 bps, minimum delay of 0.00918s, and maximum Packet delivery ratio (PDR) of 87.98%, when compared with the existing methods.

Keywords: Mobile sink, wireless sensor network, fractional concept, rider optimization algorithm, routing.

Received October 14, 2019; accept January 13, 2021

Classification of Legislations using Deep Learning

Sameerchand Pudaruth¹, Sunjiv Soyjaudah², and Rajendra Gunputh³

¹ICT Department, University of Mauritius, Mauritius

²Soyjaudah Chambers, Mauritius

³Law Department, University of Mauritius, Mauritius

Abstract: Laws are often developed in a piecemeal approach and many provisions of similar nature are often found in different legislations. Therefore, there is a need to classify legislations into various legal topics to help legal professionals in their daily activities. In this study, we have experimented with various deep learning architectures for the automatic classification of 490 legislations from the Republic of Mauritius into 30 categories. Our results demonstrate that a Deep Neural Network (DNN) with three hidden layers delivered the best performance compared with other architectures such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). A mean classification accuracy of 60.9% was achieved using DNN, 56.5% for CNN and 33.7% for Long Short-Term Memory (LSTM). Comparisons were also made with traditional machine learning classifiers such as support vector machines and decision trees and it was found that the performance of DNN was superior, by at least 10%, in all runs. Both general pre-trained word embeddings such as Word2vec and domain-specific word embeddings such as Law2vec were used in combination with the above deep learning architectures but Word2vec had the best performance. To our knowledge, this is the first application of deep learning in the categorisation of legislations.

Keywords: Deep learning, neural networks, classification, legislations.

Received October 17, 2019; accepted February 9, 2021

Two-Level Classification in Determining the Age and Gender Group of a Speaker

Ergün Yücesoy

Vocational School of Technical Sciences, Ordu University, Turkey

Abstract: In this study, the classification of the speakers according to age and gender was discussed. Age and gender classes were first examined separately, and then by combining these classes a classification with a total of 7 classes was made. Speech signals represented by Mel-Frequency Cepstral Coefficients (MFCC) and delta parameters were converted into Gaussian Mixture Model (GMM) mean supervectors and classified with a Support Vector Machine (SVM). While the GMM mean supervectors were formed according to the Maximum-A-Posteriori (MAP) adaptive GMM-Universal Background Model (UBM) configuration, the number of components was changed from 16 to 512, and the optimum number of components was decided. Gender classification accuracy of the system developed using aGender dataset was measured as 99.02% for two classes and 92.58% for three classes and age group classification accuracy was measured as 67.03% for female and 63.79% for male. In the classification of age and gender classes together in one step, an accuracy of 61.46% was obtained. In the study, a two-level approach was proposed for classifying age and gender classes together. According to this approach, the speakers were first divided into three classes as child, male and female, then males and females were classified according to their age groups and thus a 7-class classification was realized. This two-level approach was increased the accuracy of the classification in all other cases except when 32-component GMMs were used. While the highest improvement of 2.45% was achieved with 64 component GMMs, an improvement of 0.79 was achieved with 256 component GMMs.

Keywords: GMM, mean supervector, speaker age and gender classification, SVM, two level classification.

Experimental Modeling of the Residual Energy of a Rechargeable Battery-Powered Node in Wireless Networks

Naseeruddin and Venkanagouda Patil

Department of Electronics and Communication Engineering, Ballari Institute of Technology and Management Visvesvaraya Technological University, India

Abstract: This paper proposes an effective method for experimental modeling of the remaining energy in terms of State Of the Charge (SOC) of a battery-powered node in a wireless network. The SOC of a battery is used to accurately determine the remaining energy of the battery. For experimentation, three practical applications (i.e., loads) were allowed to run on the Ni-MH rechargeable battery. The real-time variations in the battery terminal voltage are captured using IC INA219 fuel gauge and an empirical equation is derived from this captured data for each application. These empirical equations are used on a node as a programmable model to experimentally verify the SOC of the application discharge curves. The developed model randomly runs the application for a random duration of time and then computes the SOC of the node. The effectiveness of the randomness in the developed model has been analyzed and found to be practically worth. The proposed work can be scaled up to any number of nodes in a wireless network. This work can benefit the researchers and the academicians working in the area of wireless networks.

Keywords: Node modelling, state of the charge, fuel gauge, wireless mobile networks, discharge-curves.

Received December 14, 2019; accepted October 25, 2020

Headnote Prediction Using Machine Learning

Sarmad Mahar¹, Sahar Zafar², and Kamran Nishat¹

¹CoCIS, PAF-Karachi Institute of Economics and Technology, Pakistan

²Computer Science, Sindh Madressatul Islam University, Pakistan

Abstract: Headnotes are the precise explanation and summary of legal points in an issued judgment. Law journals hire experienced lawyers to write these headnotes. These headnotes help the reader quickly determine the issue discussed in the case. Headnotes comprise two parts. The first part comprises the topic discussed in the judgment, and the second part contains a summary of that judgment. In this thesis, we design, develop and evaluate headnote prediction using machine learning, without involving human involvement. We divided this task into a two steps process. In the first step, we predict law points used in the judgment by using text classification algorithms. The second step generates a summary of the judgment using text summarization techniques. To achieve this task, we created a Databank by extracting data from different law sources in Pakistan. We labelled training data generated based on Pakistan law websites. We tested different feature extraction methods on judiciary data to improve our system. Using these feature extraction methods, we developed a dictionary of terminology for ease of reference and utility. Our approach achieves 65% accuracy by using Linear Support Vector Classification with tri-gram and without stemmer. Using active learning our system can continuously improve the accuracy with the increased labelled examples provided by the users of the system.

Keywords: Judgment summary, head-note prediction, machine learning, text summarization.

Received March 6, 2020; accepted September 17, 2020

Glaucoma Detection using Tetragonal Local Octa Patterns and SVM from Retinal Images

Marriam Nawaz, Tahira Nazir, and Momina Masood

Department of Computer Science, University of Engineering and Technology, Pakistan

Abstract: Glaucoma is a fatal disease caused by the imbalance of intraocular pressure inside the eye which can result in lifetime blindness of the victim. Efficient screening systems require experts to manually analyze the images to recognize the disease. However, the challenging nature of the screening method and lack of trained human resources, effective screening-oriented treatment is an expensive task. The automated systems are trying to cope with these challenges; however, these methods are not generalized well to large datasets and real-world scenarios. Therefore, we have introduced an automated glaucoma detection system by employing the concept of the Content-Based Image Retrieval (CBIR) domain. The Tetragonal Local Octa Pattern (T-LOP) is used for features computation which is employed to train the SVM classifier to show the technique significance. We have evaluated our method over challenging datasets namely, Online Retinal Fundus Image (ORIGA) and High-Resolution Fundus (HRF). Both the qualitative and quantitative results show that our technique outperforms the latest approaches due to the effective localization power of T-LOP as it computes the anatomy independent features and ability of Support Vector Machine (SVM) to deal with over-fitted training data. Therefore, the presented technique can play an important role in the automated recognition of glaucoma lesions and can be applied to other medical diseases as well.

Keywords: Retinal images, glaucoma, SVM, classification.

Received April 28, 2020; accepted November 24, 2020

Traffic-Aware Clustering Scheme for MANET

Using Modified Elephant Herding Optimization

Algorithm

Sreekanth Ramakrishnan, Latha Sevalaiappan, and Suganthe Ravichandran

Department of Computer Science and Engineering, Kongu Engineering College, India

Abstract: Clustering is the prevalent routing method in the large-scale Mobile Ad Hoc Network (MANET). The Cluster-Heads (CHs) play an important role in routing as it is transient through all communications of its associated nodes. To ensure fairness in the use of energy in all clusters, each CH has to deal with same amount of traffic. The previous clustering methods focused mainly on the distribution of equal member nodes in each cluster. They failed to consider every cluster's traffic generated. This paper introduces a novel technique for MANET clustering with Modified Elephant Herding Optimization based on the traffic generated within each cluster. This Traffic-Aware Clustering with Modified Elephants Herding Optimization (TAC-MEHO) produces optimized clusters for stable communication and is experimentally tested with well-known clustering techniques. Assessment metrics such as number of Cluster-Heads (CHs), lifetime of the network, and re-clustering rates are measured using various parameter values such as network size, network traffic and transmission distance. The results show that proposed TAC-MEHO improves the re-clustering rate by 91% and 58% when compared with Weighted Clustering Algorithm (WCA) and WCA-GA respectively. Further, it improves the network lifetime by 89% and 88 % over WCA and WCA-GA respectively.

Keywords: Clustering, MANET, traffic-aware, elephant heard optimization, cluster-head, large-scale network.

Received June 10, 2020; accepted February 25, 2021

A Novel Method for Gender and Age Detection

Based on EEG Brain Signals

Haitham Issa¹, Sali Issa², and Wahab Shah³

¹Department of Electrical Engineering, Zarqa University, Jordan

²Department of Computer Engineering, Hubei University of Education, China

³Department of Electrical Engineering, Namal University, Pakistan

Keywords: EEG, Gender, Age, CWT, CNN.

Received September 27, 2020; accepted February 9, 2021

Joint Optimization Offloading Strategy of

Execution Time and Energy Consumption

of Mobile Edge Computing

Qingzhu Wang and Xiaoyun Cui

School of Computer Science, Northeast Electric Power University, China

Abstract: As mobile devices become more and more powerful, applications generate a large number of computing tasks, and mobile devices themselves cannot meet the needs of users. This article proposes a computation offloading model in which execution units including mobile devices, edge server, and cloud server. Previous studies on joint optimization only considered tasks execution time and the energy consumption of mobile devices, and ignored the energy consumption of edge and cloud server. However, edge server and cloud server energy consumption have a significant impact on the final offloading decision. This paper comprehensively considers execution time and energy consumption of three execution units, and formulates task offloading decision as a single-objective optimization problem. Genetic algorithm with elitism preservation and random strategy is adopted to obtain optimal solution of the problem. At last, simulation experiments show that the proposed computation offloading model has lower fitness value compared with other computation offloading models.

Cost-Aware Ant Colony Optimization Based Model for Load Balancing in Cloud Computing

Malini Alagarsamy¹, Ajitha Sundarji², Aparna Arunachalapandi³, and Keerthanaa Kalyanasundaram⁴

¹Department of Computer Science, Thiagarajar College of Engineering, India

²AstraZeneca GTC, India

³Wipro Technologies Limited, India

⁴HCL Technologies, India

Abstract: Balancing the incoming data traffic across the servers is termed as Load balancing. In cloud computing, Load balancing means distributing loads across the cloud infrastructure. The performance of cloud computing depends on the different factors which include balancing the loads at the data center which increase the server utilization. Proper utilization of resources is termed as server utilization. The power consumption decreases with an increase in server utilization which in turn reduces the carbon footprint of the virtual machines at the data center. In this paper, the cost-aware ant colony optimization based load balancing model is proposed to minimize the execution time, response time and cost in a dynamic environment. This model enables to balance the load across the virtual machines in the data center and evaluate the overall performance with various load balancing models. As an average, the proposed model reduces carbon footprint by 45% than existing methods.

Keywords: Scheduling algorithms, application virtualization, power, energy.

Received August 3, 2020; accepted April 7, 2021

Spider Monkey Optimization Algorithm for

Load Balancing in Cloud Computing

Environments

Sawsan Alshattnawi and Mohammad AL-Marie

Department of Computer Science, Yarmouk University, Jordan

Abstract: Scheduling of tasks is one of the main concerns in the Cloud Computing environment. The whole system performance depends on the used scheduling algorithm. The scheduling objective is to distribute tasks between the Virtual Machines and balance the load to prevent any virtual machine from being overloaded while other is underloaded. The problem of scheduling is considered an NP-hard optimization problem. Therefore, many heuristics have been proposed to solve this problem up to now. In this paper, we propose a new Spider Monkeys algorithm for load balancing called Spider Monkey Optimization Inspired Load Balancing (SMO-LB) based on mimicking the foraging behavior of Spider Monkeys. It aims to balance the load among virtual machines to increase the performance by reducing makespan and response time. Experimental results show that our proposed method reduces tasks' average response time to 10.7 seconds compared to 24.6 and 30.8 seconds for Round Robin and Throttled methods respectively. Also, the makespan was reduced to 21.5 seconds compared to 35.5 and 53.0 seconds for Round Robin and Throttled methods respectively.

Keywords: Cloud computing, load balancing, metaheuristic optimization, spider monkeys optimization, tasks scheduling.

Received April 1, 2020; accepted January 6, 2021

Abstract: We use images in several important areas such as military, health, security, and science. Images can be distorted during the capturing, recording, processing, and storing. Image quality metrics are the techniques to measure the quality and quality accuracy level of the images and videos. Most of the quality measurement algorithms does not affect by small distortions in the image. Magnetic Resonance Imaging (MRI), Computed Tomography (CT), and Ultrasonic Imaging (UI) are widely used in the health sector. Because of several reasons it might be artifacts in the medical images. Doctor decisions might be affected by these image artifacts. Image quality measurement is an important and challenging area to work on. There are several metrics that have been done in the literature such as mean square error, peak signal-noise ratio, gradient similarity measure, structural similarity index, and universal image quality. Patient information can be an embedded corner of the medical image as a watermark. Watermark can be considered one of the image distortions types. The most common objective evaluation algorithms are simple pixel based which are very unreliable, resulting in poor correlation with the human visual system. In this work, we proposed a new image quality metric which is a Measure of Singular Value Decomposition (M-SVD). Experimental results show that novel M-SVD algorithm gives very promising results against Peak Signal to Noise Ratio (PSNR), the Mean Square Error (MSE), Structural Similarity Index Measures (SSIM), and 3.4. Universal Image Quality (UIQ) assessments in watermarked and distorted images such as histogram equalization, JPEG compression, Gamma Correction, Gaussian Noise, Image Denoising, and Contrast Change.

Keywords: Image quality measurement, M-SVD, image distortion, watermarking, medical image.

Received September 7, 2020; accepted February 1, 2021