"Annals. Computer Science Series" Journal
Romania, 300559 Timişoara, 6 Lascăr Catargiu str.
Phone: 004 0256 220 687 | Fax: 004 0256 220 690
E-mail: conference.fcia [@] tibiscus [.] ro | conference.fcia [@] gmail [.] com




During August 2019, we are on a summer vacation. We are working to update the current volume but at low speed. Sorry for the inconvenience!


Web design:
Ovidiu Crista
Updating:
Tiberiu Marius Karnyanszky
Editor-in-Chief

Annals. Computer Science Series
Tome 17, Fasc. 1


ISSN: 1583-7165 (printed journal)
ISSN: 2065-7471 (e-journal)
Owner: "Tibiscus" University of Timişoara, România
Editor: Mirton Publishing House of Timişoara, România
Distributor: "Tibiscus" University of Timişoara, România
Appearance: 30th of June, 2019

» Parameter Estimation of Cobb Douglas Production Function with Multiplicative and Additive Errors using the Frequentist and Bayesian Approaches

» J.O. Iyaniwura, A. Adedayo Adepoju and Oluwaseun A. Adesina

ABSTRACT:

Nonlinear Models are generally classified as intrinsically nonlinear and intrinsically linear based on the specification of the errors. This study was aimed at estimating the parameters of Cobb-Douglas production function with additive and multiplicative errors using the classical and Bayesian approaches. The classical nonlinear method considered is the Gauss-Newton iterative Method while the Bayesian estimation was carried out using the Metropolis-within-Gibbs with independent normal-Gamma prior. For the classical, the results showed that the estimates of the parameters of the Cobb-Douglas function with additive errors performed better than those for the multiplicative errors. However, similar estimates were obtained for both multiplicative and additive errors for the Bayesian approach. Overall, the Bayesian method performed better than the classical approach.

KEYWORDS:

Cobb-Douglas Production function, Gauss-Newton Method, Normal-Gamma Prior, MCMC.

» Received: 10 March 2018
» Accepted: 30 September 2018

9 - 15

imag/pdf.ico.gif PDF

» A Modified Ratio Techniques in Successive Sampling on Two Occasions


» K. O. Adetunji and A. O. Adejumo

ABSTRACT:

Successive sampling is a known technique that can be used in longitudinal surveys to estimate population parameter and measurement of change or sum of study variables and auxiliary variables. However, the problem of ratio estimation is incorporated to successive sampling on the samples selected over two occasions has been considered. This study aimed at proposed Ratio Estimator technique in the context of devising efficient sampling strategies for estimators. The three estimators used as the instrument for parameter estimation in the successive sampling are Simple Estimator (SEst), Linear Estimator (LEst) and proposed Ratio Estimator (PRATEst). The proposed Ratio Estimator procedure was obtained through the properties of biasness, mean square Error (MSE), minimum mean square Errors (MMSE) and Efficiency comparison (EC). Two sets of Real life data were used in this study. The first data was collected from National Population Commission (NPC) of the census conducted in Nigeria for year 1991 and 2006, also the second data set was collected from ministry of education and human capital development on the teachers and students enrolment for 2013/2014 and 2016/2017 academic sessions. In, conclusion, it is clear and also visible that proposed estimator (PRATEst) μ_PRATE=(˥x)/(˥y)*(˥z) is highly rewarding than conventional ones μ_SE=(˥x),(˥y) and (˥z) and μ_LE=(˥x)(˥y),(˥x)(˥z) and (˥y)(˥z) therefore, the proposed estimator is recommended for use in successive sampling scheme.

KEYWORDS:

Successive sampling, Ratio Estimator, Auxiliary Variable, Census, Students Enrolment.

» Received: 22 October 2018
» Accepted: 14 January 2019

16 - 25

imag/pdf.ico.gif PDF

» Towards Refining Unrated and Uninterested Items for Effective Collaborative Filtering Recommendations


» Abba Almu, Abubakar Roko, Aminu Mohammed and Ibrahim Sa’idu

ABSTRACT:

Collaborative filtering recommender systems being the most successful and widely used plays an important role in providing suggestions or recommendations to users for the items of interest. However, many of these systems recommend items to individual users based on ratings which may not be possible if they are not sufficient due to the following problems: it may lead to the prediction of uninterested popular items already known to the users because of the penalty function employed to punish those items, the sparsity of the user-item rating matrix increases making it difficult to provide accurate recommendations and also it ignores the users general preferences on the recommended items whether they are of interest to users or not. Therefore, many times uninterested items can be found in the recommended lists of an individual user. This will make user to lose interest in the recommendations if these uninterested predicted items always appear in the lists. In this paper, we proposed a collaborative filtering recommendations refinement framework that combines the solutions to these three identified problems. The framework incorporates a popularise similarity function to reduce the influence of popular items during recommendations, an algorithm to fill up the missing ratings of unwanted recommendations in the user-item rating matrix thereby reducing the sparsity problem and finally an algorithm to solicit for user feedback on the recommended items to minimise uninterested recommendations.

KEYWORDS:

Collaborative Filtering Recommendations, Sparsity, User Preference, Popular items, Uninterested Recommended List.

» Received: 8 November 2018
» Accepted: 14 January 2019

26 - 32

imag/pdf.ico.gif PDF

» Evaluation of an Optical Character Recognition Model for Yoruba Text


» Abimbola Akintola, Tunji Ibiyemi and Amos Bajeh

ABSTRACT:

The optical character recognition (OCR) for different languages has been developed and in use with diverse applications over the years. The development of OCR enables the digitization of paper document that would have been neglected over a period of time as well as serving as a form of backup for those documents. The system proposed is for isolated characters of Yoruba language. Yoruba language is a tonal language that carries accent on the vowel alphabets. The process used involves image gray scal, binarization, de-skew, and segmentation. Thus, the OCR enable the system read the images and convert them to text data. The proposed model was evaluated using the information retrieval metrics: Precision and Recall. Results showed a significant performance with a recall of 100% in the sample document used, and precision results that varies between 76%, 97%, and 100% in the sample document.

KEYWORDS:

Recognition, binarization, image digitization, accuracy, Yoruba language.

» Received: 19 October 2018
» Accepted: 14 January 2019

33 - 42

imag/pdf.ico.gif PDF

» Balanced Incomplete Sequence Crossover Design for First Order Residual Effect


» Lanre Adebara and Babatunde Lateef Adeleke

ABSTRACT:

Crossover designs are of the type in which sequence of different treatments are administered one at a time over a certain period of time, during which the presence of residual effect or rather carryover effect can longer be ignored. This paper therefore presents two first order residual effect, that is effect on the immediate next period after the period of treatment application. The first method constructs designs for any number of treatment, v for any prime number that has x=2 as primitive root of the associated Galois field. using the two algorithms I1 = ( x0, xm , x2m,..., x2k-x)and I2 = (x1, xm+1, x2m+1,..., x2k-x+1). The second method is also for the construction of designs for any number treatment, v for any prime number that has x=3 as the primitive root of the corresponding Galois field with two algorithms I1 = (x0, xm, x2m,..., x2k+m-x-1) and I2 = (x1, xm+1, x2m+1,..., x2k+m-x). By exploiting cyclic development of v-1 initial sequences of treatment order, universally optimal balanced crossover design of the first order for the two methods were constructed in this paper and the two methods generate every non-zero elements of any prime number.

KEYWORDS:

Crossover Design; Galois Field; Balanced Design; Initial Sequence; Cyclic Development.

» Received: 12 November 2018
» Accepted: 25 January 2019

43 - 46

imag/pdf.ico.gif PDF

» Survival Analysis of Breast Cancer in Nigeria


» Osuolale Peter Popoola, Ayomide Oluwaghenga and Abosede Titilope Popoola

ABSTRACT:

In Nigeria as well as in other parts of the world, Breast cancer remained the most common cancer among women and the second leading cause of death. This research attempts to carry out data analysis on breast cancer data in Nigeria so as to determine the trend movement of the disease, determine how age influences the survival of patient, test for the significance in the distribution of survival time of the patients, measure the average survival time of the patients after treatment, estimate the time to event of interest (death) and estimate the probability of survival. The results of various data analysis showed that the median survival time until the event occurs is 15 days for male and 13 days for female. Kaplan-Meier Estimator shows that female response to treatment faster than male patients with the mean time of 24.8 days and t 21.90 days respectively, the survival plot shows that the probabilities of surviving is decreasing as time progresses, it also revealed that in all age groups, there is a decrease in their chance of surviving. The result of cox proportional hazard regression analysis shows that Age and Length have high statistical significant coefficients which means that the risk of death is higher in Year with a positive value while the risk of death is low in Age, Sex, and Length spent in hospital with negative value of coefficient. The Hazard Ratio (HR) of 0.993186, indicates a strong relationship between the patients’ Age and patients risk of death and between the time spent in the hospital and decrease risk of death. The beta coefficient for sex gives -0.1828 indicates that female patients have lower risk of death than male patients and the hazard ratio gives 0.832338. The Wilcoxon test and Long rank test revealed that there are no statistically differences in the survival rates between males and females but there is a statistically differences in the survival rates between Dead and Alive. It also shows that female patients have lower risk of death than male patients with the beta coefficient of -0.1828 while hazard ratio gives 0.832338.

KEYWORDS:

Survival Analysis, Breast Cancer, Kapler Meier Estimator, Log-Rank Test and Wilcoxon Rank Test and Cox Proportion Regression.

» Received: 5 December 2018
» Accepted: 29 January 2019

47 - 56

imag/pdf.ico.gif PDF

» Differential Equation as a Solution for Population Problems


» Y. Zakari and A. Hassan

ABSTRACT:

World of mathematical ideas, where a model is established, we then manipulate the model by the help of already existed techniques or computer numerical computations, finally we then investigate the model that will lead us to the solution to our mathematical problems related to the model, which is translated into a real life applications. The application of first order differential equation in Growth and Decay problems will study the method of variable separable and the model of Malthus (Malthusian population model), where we use the methods to find the solution to the population problems which are of use in mathematics, physics and biology especially in dealing with problems involving growth and decay problems that requires the use of Malthus model.

KEYWORDS:

Malthus model, Differential equation, Growth rate, Decay rate, Population rate, Population, Order, Attenuation.

» Received: 19 January 2018
» Accepted: 7 February 2019

57 - 62

imag/pdf.ico.gif PDF

» Arabic Digital Resources in the Service of Learning


» Ali Sadiqui, Ahmed Zinedine and Mohamed El Hari

ABSTRACT:

Video-pedagogy knows a growing popularity, due to its many supposed virtues. Compared to static documents, the animations have the power to attract attention and increase the interest and motivation of the reader. Un-fortunately, the resources available on the Internet in Arabic language in this area do not reflect its position as a language used by more than 300 million people. Our aim in this contribution is to present a project that was designed and steered in this context. This project is motivated by the desire to integrate vid-eo animations as educational support for learners of the OFPPT (Office de Formation Professionnelle et de Promotion du Travail) as well as to contribute to enrich the Arabic content on the Internet by quality products in this area. We then, in collaboration with a team of practitioners and with the participation of some learners designed and tested a series of video animation in Arabic language allowing the acquisition of theoretical concepts taught in the classroom. The final product has positively impacted the assimilation of courses. This contribution will return to all the components of this project.

KEYWORDS:

Arabic language, Arabic digital resources, Video-pedagogy, reverse class.

» Received: 11 November 2018
» Accepted: 25 January 2019

63 - 68

imag/pdf.ico.gif PDF

» Clonal Selection Algorithm for Feature Level Fusion of Multibiometric Systems


» Oluyinka T. Adedeji, Adeleye S. Falohun, Oluwaseun M. Alade, Elijah O. Omidiora and Stephen O. Olabiyisi

ABSTRACT:

Multimodal biometric makes use of two or more biometric modalities to overcome some of the limitations of unimodal biometric system. Feature level fusion has been shown to provide a more secured recognition system with higher performance accuracy. However, associated with feature level fusion is the problem of high dimensionality of the combined feature, therefore in this paper, Discrete Wavelet Transform (DWT) is used for feature extraction while fusion is performed at the feature selection phase using Clonal Selection Algorithm (CSA). The performances of the bimodal systems indicate increase in recognition accuracy compared to their unimodal counterparts.

KEYWORDS:

Biometrics, Fusion, Multimodal, Discrete Wavelet Transform, Clonal Selection Algorithm.

» Received: 11 October 2018
» Accepted: 26 November 2018

69 - 75

imag/pdf.ico.gif PDF

» Development of a Modularized Model for a Multi-Process Food Manufacturing System using Hierarchical Timed Coloured Petri Nets


» Rafiu A. Ganiyu, Oladotun O. Okediran, Busirat O. Muraina and Taofeeq A. Badmus

ABSTRACT:

Concerted efforts are being made yearly by researchers to improve food manufacturing systems through modelling in order to meet current and future market demands. However, most existing empirical studies are limited to development of non-modular models for multi-process food manufacturing systems. Hence, in this paper, a Hierarchical Timed Coloured Petri Net model, which could accommodate future modifications, was developed for Gari manufacturing system. The model comprises Peeling, Washing, Grating, Fermentation, Dewatering, Pulverization, Garifying, Cooling, Sifting and Packaging modules. The model was simulated to determine the average times used and utilized-resources rates for productions of 200, 400 and 600 kg of Gari, and to study what could happen to production times in the possibility of additional resources for productions of 800 and 1000 kg of Gari. The model was validated at 5% significant level using statistical t-Test. The model could be efficiently utilized to study and improve Gari production working plan.

KEYWORDS:

Petri Nets, Food, Gari production, Module, Process, Model.

» Received: 26 October 2018
» Accepted: 14 January 2019

76 - 86

imag/pdf.ico.gif PDF

» Time Series Analysis of Brent Crude Oil Prices Per Barrel: A Box-Jenkins Approach


» Adebowale Olusola Adejumo and James Daniel

ABSTRACT:

This study was aimed at analyzing the Brent Crude Oil Price Per Barrel. The data was extracted from the Organisation of Petroleum Exporting Countries (OPEC) Bulletin and it covered the period of January 1985 to September 2016. The Box and Jenkins approach of model identification, parameter estimation and diagnostic checking was adopted in the analysis with the aid of S-plus Package. In view of the above problem, this research was set to: explore and explain the behavior of the series; determine the best model; and forecast future values of the series. With the aid of S-Plus programming ware, R language ware; ARIMA (2,1,2) (2,0,0) [12] was reached as an optimum and parsimonious model for the series. An arithmetic increase in the price was forecast.

KEYWORDS:

Autocorrelation, Partial Autocorrelation, Autoregressive Integrated Moving Average (ARIMA), Autoregressive Moving Average (ARMA), Autoregressive (AR), Moving Average (MA), Differencing, Brent Crude Oil, Stationary Series.

» Received: 27 January 2018
» Accepted: 6 April 2018

87 - 101

imag/pdf.ico.gif PDF

» Performance Evaluation of a Cloud-based Picture Archiving and Communication System (PACS)


» Jinmisayo Awokola, Justice Emuoyibofarhe, Christoph Meinel and Funmi Ajala

ABSTRACT:

In the field of medical informatics, there has been a paradigm shift from the adoption of the conventional Picture Archiving and Communication System (PACS) to cloud-based PACS because of the cost-effectiveness and storage computational efficiency. However, cloud-plugged technologies are subject to security and network performance issues because of the internetworking and interoperability over the internet. Hence, this work focused on the development and evaluation of a secured cloud-based PACS framework so as to assess the usability of the technology in developing countries of the World through a case study approach.

KEYWORDS:

.

» Received: 19 August 2018
» Accepted: 3 March 2019

102 - 108

imag/pdf.ico.gif PDF

» Facial Recognition System: A Shift in Students Attendance Management


» Kehinde Sotonwa and Oluwashina Oyeniran

ABSTRACT:

It was observed that the habitual method of taking attendance easily allow impersonation and time consuming. In the global community, various facial recognition based attendance frameworks have being proposed and successfully implemented, many of which are of one limitation or the other. Hence, this study proposed a multi-algorithm approach to design and implement an automated and reliable system using biometric for effective attendance taking in Nigerian institutions. The system employed bimodal biometric technique for combining Principal Component Analysis (PCA) with Histogram of Oriented Gradient (HOG) for feature extraction and Artificial Neural Network (ANN) combined with Support Vector Machine (SVM) for classification. For the design and implementation, the system made use of the in-built webcam to capture its images for enrolment and verification. The system was trained with five hundred registered and fifty unregistered facial images.

KEYWORDS:

PCA, HOG, ANN, SVM, C#, SQL.

» Received: 21 January 2019
» Accepted: 3 March 2019

109 - 113

imag/pdf.ico.gif PDF

» An Enhanced Round Robin Virtual Machine Load Balancer for Cloud Infrastructure


» Timothy Moses, Edward O. Agu, Okwori Anthony Okpe and John A. Oladunjoye

ABSTRACT:

With continuous increase in volume of data and its computational capacity in cloud, support for users request in cloud cannot be ignored. Considering problem of load imbalance resulting in high throughput and response time, there is the need therefore to develop a better model that will improve resource utility and performance of distributed systems so as to reduce response and processing time while providing better Virtual Machine (VM) efficiency. This paper analysed Round Robin load balancing algorithm and the deficiencies of this algorithm served as a basis for improvement in the proposed load balancer. The proposed algorithm has three phases; VM categorization phase that re-arrange VMs in the increasing order of their processing speed with 80% of total threshold value of each VM computed. Allocation phase helps find suitable VM for execution of cloudlet and the reliability assessment phase ensures that each VM performs at it optimal level before its consideration for allocation of cloudlets. An extensive simulation was carried out to evaluate the proposed algorithm using cloud analyst simulator. Results obtained from simulation shows that the proposed algorithm yields better response time, datacentre processing time and a lower turnaround time than the existing round robin load balancing algorithm.

KEYWORDS:

Virtual machine, datacentre, virtualization, threshold value, cloudlet, user base, turnaround time, load balancer.

» Received: 9 January 2019
» Accepted: 3 March 2019

114 - 121

imag/pdf.ico.gif PDF

» Construction of Steiner Triple System STS(2n +1) from a Class of Pairwise Balanced Designs


» Osuolale Peter Popoola, Benjamin A. Oyejola, Ayanniyi Ayanrinde and Matthew Odusina

ABSTRACT:

Pairwise Balanced Design (PBD) is a pair (X, Ɓ) where X is a set of treatments and Ɓ is a collection of subsets of X called blocks, such that each pair of treatments is contained in precisely one block. PBD plays important role in design theories, it is used to construct other important designs such as Steiner Triple System (STS). A Steiner triple system is an ordered pair (X, Ɓ), where X is a finite set of points (Treatments) and B is a set of all 3-element subsets of X called triples, such that each pair of distinct elements of X occurs together in exactly one triple of B. The research work aims at applying a class of PBD(n, K, 1) when K ={3, 4} and λ = 1 to construct STS(2n +1). Theorem was proposed and proved and a certain inequality was derived as condition which must be satisfied for the construction to hold. Thus, for all n ≡ 1, 3(mod 3) of any PBD(n, {3, 4}) there exists an STS(2n +1) provided n ≥ l(s -1) + 1, where l is the size of the largest block of the PBD and s is the size of the smallest block of the PBD. Hence, STS(21) was constructed from a PBD(10, {3, 4}, 1).

KEYWORDS:

Block Designs, Pairwise Balanced Design (PBD), Steiner Triple System (STS) and Combinatorial System.

» Received: 14 January 2019
» Accepted: 13 March 2019

122 - 126

imag/pdf.ico.gif PDF

» Construction of Congruent Classes of Pairwise Balanced Designs using Lotto Designs


» Osuolale Peter Popoola and Benjamin A. Oyejola

ABSTRACT:

Among the Incomplete Block Designs(IBDs), Balanced Incomplete Block Designs (BIBDs) are mostly studied. However, BIBD is not available for all parameters of most designs therefore limited in applications. So, in place of BIBD there is need for another incomplete block design that could be used for varieties of applications. Thus, the need for Pairwise Balanced Designs (PBDs). PBD (n, K, λ) is a block design where n is the number of treatments, where K= {k1, k2…, kb} is the set of sizes of block and λ is number of time a pair of treatment appear together within blocks. Also, little is known about the construction of PBDs using Lotto Designs (LDs). An LD(n, k, p t) is a set of k-blocks of an n-treatments such that any p-sets intersect at least one k-block in t number of times. The aim of the study is to provide a simple method for constructing two classes of PBD(n, K, λ) when = K {3, 4} or {3, 4, 5} using appropriate LDs; establish conditions for the identification of LDs that could use to construct the classes of PBDs; and derive theorems and simple steps for the construction of PBDs from LDs. The research work utilized the Li’s inequality to obtain LDs that are PBDs using r and λ obtained from the classes of PBDs on the Li inequality. Some LDs were generated and based on the structure of the classes of the PBDs to be constructed, some conditions were imposed. Hence some LDs were found to qualify as PBDs. Theorems were proposed and proved. Hence, the following results were obtained: Any LD(n, k, p, t) satisfying the conditions: 3 ≤ k ≤ n and n = p qualify as PBDs and the proposed theorems were: (a) 2-LDs(n, k, p, t)(n, k+1, p, t) is a PBD(n, {3, 4}) if and only if n ≡ 0, 1 (mod 3); (b). 3-LDs(n, k, p, t)(n, k+1, p, t)(n, k+2, p, t) is a PBD(n, {3, 4, 5}) if and only if n ≡ 2, 3 (mod 4). Thus, two classes of PBDs(n, K, λ) when K= {3, 4}) and K= {3, 4, 5} were constructed from the appropriate LDs for all admissible n-treatments satisfying the specified conditions.

KEYWORDS:

Block Designs, Incomplete Block Designs (IBD), Balanced Incomplete Block Designs (BIBDs), Pairwise Balanced Designs (PBDs) and Lotto Designs (LDs).

» Received: 12 January 2019
» Accepted: 3 March 2019

127 - 134

imag/pdf.ico.gif PDF

» Adaptive Particle Swarm Optimization Scheduling Model for Jobs on Cloud


» Alaba T. Owoseni, Olatunde Iyaniwura and Lukman Adebayo Ogundele

ABSTRACT:

Cloud computing is a special network-based computation that makes computational services available to customers on demand. In this computational paradigm, resources are managed by organizations that have all it takes to purchase and manage them since they are cost intensive and as such, the needy of computer-based services pay the organizations as virtualized resources are used to provide services to the needy. On cloud, effective and efficient computation depends on the technique used in scheduling tasks among computational resources especially the virtual machines and this has called for many researches from various quarters. In this paper, we propose an improved particle swarm optimization algorithm whose parameters (cognitive and social coefficients and inertial weight) are adaptive as a measure against premature convergence of the algorithm. CloudSim simulator was used to simulate the operation of the algorithm and the performance was compared with classic particle swarm optimization method for better result.

KEYWORDS:

Adaptive particle swarm optimization, inertial weight, cloud computing, cognitive coefficient, social coefficient, makespan.

» Received: 15 January 2019
» Accepted: 13 March 2019

135 - 141

imag/pdf.ico.gif PDF

» Improved Variance Estimator using Linear Combination of Tri-mean and Quartile Average


» J.O. Muili, A. Audu, R. V. K. Singh, A.B. Odeyale and A. Adebiyi

ABSTRACT:

In this paper, we proposed a ratio cum product estimator for estimating finite population variance. The proposed estimator was obtained by transforming [MS17] estimator. The bias and mean square error (MSE) of the proposed estimator has been obtained and the conditions for it efficiency over some existing variance estimators have been established. The efficiency of proposed estimator based on optimal value of the constant, exhibit significant improvement over the estimators considered in the study. The numerical illustration was also conducted to corroborate the theoretical results. The results of the empirical study show that the proposed estimator is more efficient over existing estimators.

KEYWORDS:

Tri-mean, Quartile, Mean Square Error, Efficiency, Variance.

» Received: 14 December 2018
» Accepted: 29 January 2019

142 - 147

imag/pdf.ico.gif PDF

» Improved Estimators of Finite Population Variance using Unknown Weight of Auxiliary Variable


» J. O. Muili, A. Audu, R. V. K. Singh and I. A. Yunusa

ABSTRACT:

In this paper, we proposed a class of ratio estimators for finite population variance. The proposed estimators were obtained by transforming [S+18] estimators. The properties (Bias and Mean Square Errors (MSEs)) of the proposed estimators have been obtained and the conditions for their efficiencies over some existing variance estimators have been established. The efficiencies of proposed estimators based on optimal value of the constant, exhibit significant improvement over the estimators considered in the study. The numerical illustration was also conducted to corroborate the theoretical results. The results of the empirical study show that the proposed estimators are more efficient over existing estimators.

KEYWORDS:

Efficiency, Auxiliary variable, Median, Mean Square Error, Ratio Estimator.

» Received: 14 December 2018
» Accepted: 29 January 2019

148 - 153

imag/pdf.ico.gif PDF

» Numerical Solutions to the Extinction Problem of Polio Transmission Agents in Nigeria


» Johnson Gbola Olawuwo and Olabisi O. Ugbebor

ABSTRACT:

In spite of local and international efforts to end the scourge of poliomyelitis in the world, the disease still exists around the world with new cases being detected in otherwise polio-free areas while the known endemic places of Asia and Sub-Saharan Africa are yet to be ridden of the epidemic. Knowing with accuracy whether the disease has finally ended in a population has always being a subject of interest to mathematicians, statisticians and epidemiologists. In this study, we have defined a conditional probability function xic(t) as the probability of zero transmission agents given that there are initially (i) symptomatic (ii) infectives and (iii) asymptomatic carriers. Using the properties of Poisson process, two backward equations were derived which were jointly solved numerically . With these numerical solutions, we were able to formulate a model for the two identified transmission agents of poliomyelitis.

KEYWORDS:

Modelling, Backward Equation, Picard’s Method, Polio-free Period, Polio Virus.

» Received: 13 February 2019
» Accepted: 24 March 2019

154 - 156

imag/pdf.ico.gif PDF

» A multi-level authentication scheme for controlling access to private and sensitive information of an enterprise


» M. O. Odim, G. S. Fashoto and V. I. Nsiamuna

ABSTRACT:

This study proposed a multilevel authentication security scheme for controlling access to private and sensitive information against unauthorised users. The scheme is composed of face recognition at the first level and username/password authentication, at the other level. The face recognition was modelled using principal component analysis, while the username and password employed VB.Net password tool. One hundred users were enrolled and their faces captured using a webcam; they were afterward used to access the performance of the proposed system. The results showed that access could only be granted by successful validation of the combined authentication levels. However, it was observed that the face recognition accuracy of the scheme could be impeded by the wrong positioning of the capturing device. Nevertheless, experiment showed that the scheme could provide a stronger protection of sensitive information than the single security level authentication scheme.

KEYWORDS:

Authentication, Access control, Information security, Principal Component Analysis, Face Recognition.

» Received: 31 January 2019
» Accepted: 24 March 2019

157 - 163

imag/pdf.ico.gif PDF

» Assessment of Student’s Performance in Tertiary Education in Nigeria using Statistical Quality Control


» Olajide Johnson Taiwo and Ayansola Olufemi A.

ABSTRACT:

This paper focuses on monitoring the performance process in tertiary education in Nigeria using The Polytechnic, Ibadan as a case study to examine if the institution, students and lecturers meet set academic standard or not. The number of students taken each of the eighteen 100 level courses with their corresponding number of failure were extracted from the concerned department and the proportions of defectives (failures) were obtained. P-chart is used to monitor the failure rate using excel. The analysis was carried out using SPSS software with Control Limit (CL), Upper Control Limit (UCL) and Lower Control Limit (LCL) of 0.0802, 0.1562 and 0.0042, respectively. The chart revealed that the failure rate is out of statistical control. Based on these findings, it was noticed that the contents of some of these courses are School certificate syllabus which students must have done in secondary school but despite this; the failure rate in some courses are high. This indicates that the performances of student are not encouraging in some of the courses offered which might be caused by series of reasons.

KEYWORDS:

Control limit, failure rate, non-conformable, proportion, quality control.

» Received: 22 February 2019
» Accepted: 2 April 2019

164 - 168

imag/pdf.ico.gif PDF

» Intrusion Detection using Deep Learning Technique: A Review


» Olatunde Iyaniwura, Festus Ayetiran and Alaba T. Owoseni

ABSTRACT:

This work is a survey paper that represents the review of the current research in deep learning. Developing a flexible and efficient NIDS for unforeseen and unpredictable attacks, deep learning based provides more efficient result. Shallow learners mostly depend on the features used for creating the prediction model. On the other hand, deep learners have the potential to extract better representations from the raw data to create much better models. The self-taught learning algorithms work best when they are allowed to learn rich models using large amounts of unlabeled data (millions of examples). The STL model could be implemented in an environment where data is unclean. Though, an important consideration is to recognize that the best practice for applying any model, would be to ensure that the data is clean. Deep learning give room for the elimination of manual work done by system administrator.

KEYWORDS:

Intrusion Detection, shallow learning deep learning Self-Taught Learning Encoder.

» Received: 21 January 2019
» Accepted: 24 March 2019

169 - 178

imag/pdf.ico.gif PDF

» Robustness of Bootstrap Estimators to Influential Observations


» T. P. Ogundunmade and A. A. Adepoju

ABSTRACT:

The existence of outliers in the original sample may create problem to the classical bootstrapping estimates. There is possibility that the bootstrap samples may contain more outliers than the original dataset since the bootstrap re-sampling is with replacement. Consequently, the outliers will have an unduly effect on the classical bootstrap mean and standard deviation. This study examined the performance of three bootstrap estimators namely: Case Bootstrapping, Fixed-X Bootstrapping and Residual Resampling method under different levels of outliers. The objective was to determine which of these bootstrap methods is resistant to the presence of outliers in the data. Three levels of outliers; 5%, 10% and 20% were considered and injected into sample sizes, N = 20, 30, 50, and 100 each replicated 1000 and 5000 times respectively. The performances of the bootstrap methods were evaluated using the mean, standard error, absolute bias, mean square error and the root mean square error. The results showed that the Residual resampling Bootstrap performed better than the other two estimators.

KEYWORDS:

Bootstrapping, case bootstrapping, Fixed-X Bootstrapping.

» Received: 10 March 2019
» Accepted: 7 May 2019

179 - 188

imag/pdf.ico.gif PDF

» Reconciliation of Economic Data with Irregular Benchmarks: A New Approach


» Isaac O. Ajao, Femi J. Ayoola and Joseph O. Iyaniwura

ABSTRACT:

Benchmarking problem arises when the sum of sub-annual totals of a particular economic variable from a source are not equal to the annual totals from another source. As a result, there is a need for reconciliation of the two datasets from different sources for such data to be reliably used for research purposes. Earlier research showed that Proportional First Difference (PFD) was found to be the best method for solving benchmarking problem in indicator series when compared with other methods. In this paper, we proffered an alternative solution to the benchmarking problem. This proposed benchmarking method is based on the generalized least squares regression model which is grounded on a statistical model that allows for the presence of bias and autocorrelated errors in the indicator, and the presence of non-binding benchmarks. Using the simulated quarterly series by Dagum and Cholette (2006), the results show that the method exactly preserves the quarter-to-quarter movement, a residual adjustment is made, and the final benchmarked series converges to the bias-corrected series at the end. None of these desirable features was exhibited by the existing methods of solving benchmarking problem.

KEYWORDS:

Bias-corrected, Autocorrelated errors, Benchmarking, Growth rate, indicator.

» Received: 14 February 2019
» Accepted: 7 May 2019

189 - 194

imag/pdf.ico.gif PDF

» Effecting Secure Mutual Agreement in a Mutually-Suspicious Vicious Party


» Odule Tola John, Adesina Ademola Olusola and Solanke Olakunle O.

ABSTRACT:

In a few situations, the ability to produce legitimate assent is shared among an arrangement of clients or PCs. It isn't advantageous that a solitary individual has the learning of the mystery key and signs every one of the texts at sake of the society. This reality diminishes information security and dependability. An answer to take care of these issues is to circulate the responsibility of assenting among P={p1,p2,Λ,pn} an arrangement of 'λ players, where a monotone expanding group of Γ⊂2P qualified subsets must be characterised. In our plan, we have proposed an assent plan that is secure even within the sight of an enemy who defiles and controls the conduct of some subset of Λ⊂2P exploitative players. Our plan expects a latent foe structure that is monotone perishing in which the foe structure Λ is controlled by its premise Λ0={B∈Λ|B∪{Pi}∉Λ,for all Pi∉B}. Our proposed plan is secure both in the computation and information theoretic models.

KEYWORDS:

Threshold mystery sharing, distributed signature, information security, passive adversary model, perfect security.

» Received: 23 February 2019
» Accepted: 5 May 2019

195 - 202

imag/pdf.ico.gif PDF

» Integrating Art with STEM Education - STEAM Education in Vietnam high schools


» Tran Trung Tinh and Le Hong Quang

ABSTRACT:

STEAM stands for Science (S), Technology (T), Engineering (E), Art (A), and Mathematics (M). STEAM is essentially intended to equip learners with the necessary knowledge and skills related to the fields of science, technology, engineering, art and mathematics. These knowledge and skills must be integrated and complementary to help students not only understand the principles but also be able to practice and create products in daily life. In the context of the ongoing 4.0 industrial revolution, jobs in the social sector related to the field of STEAM education tend to increase. In this study, the author investigated 120 students from five regional high schools in Viet Nam, 45 high school teachers in order to understand the current status of STEAM education in Vietnam high schools, then give some recommendations for the development of STEAM education in high schools in Vietnam. The results are as follows: students recognized the meaning and necessity of STEAM education as a problem-solving procedure that resulted in increased STEAM literacy and a development of concepts through sharing opinions. Most students indicated that they would frequently make use of the knowledge they learned in the STEAM program in their science class because it allowed them to have a better understanding of the problem-solving process. Therefore, STEAM programs as develop creative problem-solving abilities by introducing new ideas.

KEYWORDS:

STEAM, Science, Technology, Engineering, Mathematics, Arts integration; education; creativity; 21st century skills.

» Received: 5 March 2019
» Accepted: 7 May 2019

203 - 213

imag/pdf.ico.gif PDF

» Differential Incidence of Type II Diabetes Mellitus between Male and Female using Beta Binomial Model


» Matthew Aina, Muhammed Besiru Jibrin, Usman Abubakar Jauro and Olukayode Aiyeniko

ABSTRACT:

One of the life threaten diseases with a long duration and considerable associated morbidity is type II diabetes mellitus. This disease affects a large populations around the world regardless of economic status and its predominance is rising despite the struggles of clinicians, researchers and public health professionals. This paper intends to analysis gender sensitivity to type II diabetes mellitus. Secondary data used in this study was obtained for a period of 10 years from medical records of General Hospital, Ifaki-Ekiti, Ekiti state, Nigeria. A Beta-Binomial Model was applied by incorporating Bayesian Statistical Technique to determine the hypothesis testing, credible interval and posterior probability distribution of the incidence of type II diabetes mellitus between Male and Female patients. Experimental results revealed that male patients are more sensitive to type II diabetes mellitus, therefore, males are more affected with type II diabetes mellitus than female.

KEYWORDS:

Beta-Binomial, Bayesian, Diabetes Mellitus, Morbidity, Sensitivity nature.

» Received: 18 March 2019
» Accepted: 28 May 2019

214 - 219

imag/pdf.ico.gif PDF

» Evaluation of Accuracy between Item-Based and Matrix Factorization Recommender System


» Oloyede I. and Hassan I. A.

ABSTRACT:

Recommender systems is used by e-commerce websites and streaming services to predict user opinion about products. This work examined two specific recommender algorithms, matrix factorization collaborative filtering algorithm and Item-based collaborative filtering, which utilizes item similarity. This study to compared the prediction accuracy of the algorithms using Mean Square Error and Root Mean Square Error criteria. The work yielded a results which indicated that the matrix factorization collaborative filtering algorithm is more accurate than the Item-based collaborative filtering algorithm. From the study, the results of evaluation metrics showed that matrix factorization method having RMSE of 0.916250 and MAE 0.708731 scales slightly better than the item-based method having RMSE and MAE of 0.937089 and 0.719434. The study concluded that the matrix factorization method is more accurate than the item-based method when evaluating their prediction accuracy using RMSE and MAE.

KEYWORDS:

item-based, recommendation system, matrix factorization, filtering.

» Received: 5 February 2019
» Accepted: 24 March 2019

220 - 228

imag/pdf.ico.gif PDF

» Detection of Upper Outliers in an Exponential Sample using Multiple Outlier Tests


» Dare Remi Julius, Adesina Sunday Olumide, Famurewa Kehine Dolapo and Owoseni Timothy

ABSTRACT:

Outlying values have been an issue of concern to researchers and data analysts. The study of multiple outliers is imperative because outlying values may lead to model misspecification, wrong estimation of parameters and incorrect statistical results. This paper proposed Tietjien-Moore test statistics for upper outliers in exponential samples. A simulated study was carried to investigate the strength of the test statistics.

KEYWORDS:

Outlier, Exponential distribution, Tietjen-Moore test, Gap test, Simulation.

» Received: 6 January 2019
» Accepted: 3 March 2019

229 - 234

imag/pdf.ico.gif PDF

» Mathematical Analysis of the Effect of Maternal Immunity on the Global Eradication of Measles


» Abayomi Ayoade, Agboola Sunday and Mohammed Ibrahim

ABSTRACT:

Measles susceptibility depends largely on the influx of newborns. Besides, expectant women who contract measles are at the risks of early labor, miscarriage or babies with poor weight. On that ground, we propose a new deterministic mathematical model to study the effect of inborn immunity on the global eradication of measles. The model is based on dividing the entire human population into six disjoint subpopulations. The stability theory of nonlinear differential equation is used to analyse the model. We derive the reproduction number of the model by using the next generation matrix approach and conduct numerical simulations to verify the analytical results. One of the fundamental results of the study is that global eradication of measles is not feasible if efforts are only concentrated on getting more and more pregnant women immunised without taking measures to reduce the rate of exposure of the susceptible individuals to measles infections.

KEYWORDS:

Measles, Model, Immunity, Reproduction Number, Simulation.

» Received: 13 March 2019
» Accepted: 28 May 2019

235 - 241

imag/pdf.ico.gif PDF

» Investigating methods of estimating parameters in Cox model with two incorrectly specified random effects: A simulation study


» O. I Adeniyi and B. A Oyejola

ABSTRACT:

Most observational data are found to be clustered. In situations like this, it is important to investigate whether there is variation in the predictor effect between the clusters. Such inter-cluster variation cannot be explained only by the heterogeneity of predictor effects across the cluster but also by heterogeneity of their baseline hazard risk. The aim of this work is to use the Penalized Partial likelihood method and Hierarchical likelihood to estimate the parameters of the model Cox model with two additive random effect when the random effect are wrongly specified via a simulation study for various cluster sizes, number of clusters, censoring percentages and magnitude of the random effect variance. The simulation study showed Hierarchical Likelihood estimates the random effects well than the Penalized Partial Likelihood but both methods estimates the fixed effect well.

KEYWORDS:

Clustering, random effects models, censoring, penalized partial likelihood, hierarchical likelihood.

» Received: 29 March 2019
» Accepted: 14 June 2019

242 - 251

imag/pdf.ico.gif PDF

» Performance of an Optimum Autoregressive Spectral Estimation Method Based on Signal Length


» Abass Taiwo and Timothy Olatayo

ABSTRACT:

The spectrum of a time series is the distribution of variance of the series as a function of frequency and the objective of spectral analysis is to estimate and study the spectrum. Then, this research work focused on estimation of the spectrum and efficiency of some spectral estimation methods with respect to frequency content analysis of a medical time series data (Heart-beat readings). Modified information criteria for chosen an optimal model for autoregressive spectral estimation method were proposed. The performance of the Periodogram, Thomson’s Multitaper and Autoregressive with burg spectral estimation were examined using 1000 heartbeat readings. From the results obtained, the proposed modified information criteria were used to select AR(9) as the optimal model and spectral estimates of AR(9) depicted the distinct spectral components not resolved by the Periodogram and Thomson’s Multitaper. The Thomson’s Multitaper method showed a better resolution that Periodogram but a close inspection revealed the existence of spectral leakage and it failed to resolved different spectral components along the frequencies. In conclusion, AR(9) showed a better spectral estimates and resolution along the frequencies and this made it better than Periodogram and Thomson’s Multitaper spectral estimation methods when analysis the frequency content of 1000 heartbeat readings.

KEYWORDS:

Spectrum, Autoregressive model, Heartbeat, Spectral estimation methods, Periodogram and Thomson’s Periodogram.

» Received: 31 March 2019
» Accepted: 14 June 2019

252 - 257

imag/pdf.ico.gif PDF

» Design and Analysis of a Formal Computational Ontology Model for Yoruba Naming Convention


» Omotosho L. O., Akanbi C. O., Otubu O. O. and Alade S. M.

ABSTRACT:

Yorùbá Culture is at the risk of going into annihilation as its rich culture and custom are being lost due to inadequate knowledge preservation measures to safeguarding its learning, teaching and practice. Folktale is the commonest method used in transferring knowledge and information from one generation to another generation. However, this is not possible and reliable as vital information and data would be lost in the face of current globalisation and enculturation. This study was carried out in order to protect Yorùbá Cultural heritage with a focus on Yorùbá Naming practice. In order to safeguard and protect the knowledge contained in the cultural heritage from being endangered, the need to develop a knowledge management system for Yorùbá Naming is required. Data on Yorùbá names were obtained from documented text and knowledge about the naming convention was elicited from domain and ontology experts using a combined method of observation and consultations. Moreover, Formal Concept Analysis (FCA) method was used in building the binary relations, which was used in the computational design of the Yorùbá artefact. The design was implemented and validated using an Ontology Web Language (OWL) tool called Protégé. The ontology model represents the documentation of the elicited knowledge of the YCH domain. Finally, the formal digital artefact resulting from the analysis and design of an ontology model or framework serves as a useful software artefact for formalising the concept of naming in Yorùbá culture. This research is very vital (indispensable) so as to preserve the rich culture of learning names, which can be extended to different societies and clans.

KEYWORDS:

Yorùbá Culture, Yorùbá Naming Convention, Ontology, Formal Concept Analysis, Ontology Web Language (OWL), Knowledge management, Computation artefacts.

» Received: 20 February 2019
» Accepted: 8 May 2019

258 - 267

imag/pdf.ico.gif PDF

» Comparison of Adaboost and Bagging Ensemble Method for Prediction of Heart Disease


» Yusuf Olatunde, Lawrence Omotosho and Caleb Akanbi

ABSTRACT:

The medical industry arguably generates the largest amount of data on a daily basis. Extraction of new and useful information from the bulk of data generated is very tedious. Although, it contributes to the quality of service rendered in the health sector. Data mining techniques are among the major approach that shows promising result when applied in diagnosing patient and prediction of diseases. In this study, AdaBoost and Bagging are used to support classifiers such as Naïve Bayes, Neural Network in prediction of heart disease while Random Forest was applied separately. Comparison of the experiment results focus majorly on the ensemble method used (AdaBoost and Bagging). With respect to this study, Bagging outperforms AdaBoost in term of Accuracy and other parameters such as Kappa Statistics, weighted average of ROC, Precision and MCC. It is therefore recommended as a good supportive technique for weak classifiers. Although, both Bagging and AdaBoost decline in performance when applied on rigorous dataset.

KEYWORDS:

Heart Disease, Adaptive Boosting, Bootstrap Aggregation, Neural Network, Naïve Bayes, Random Forest.

» Received: 16 February 2019
» Accepted: 28 March 2019

268 - 279

imag/pdf.ico.gif PDF


» www.fcia.tibiscus.ro |  » www.fcia.tibiscus.ro/ccsitia/ |  » www.geogebra.tibiscus.ro |  » www.tibiscus.ro |  © 2003-2019, Last update: 21.08.2019 / 18:00