Computer Science -Scholarly Publications
Permanent URI for this collection
Browse
Browsing Computer Science -Scholarly Publications by Author "Adewole, A.P."
Now showing 1 - 20 of 21
Results Per Page
Sort Options
- ItemOpen AccessClustering of large dataset using adaptive resonance theory 2 (Art 2)(Lautech Journal of engineering and Technology, 2014) Adewole, A.P.; Amuda, T.J.Each and every day we generate a vast amount of data from cloud, mobile and social technology, consequently, there is a need to make sense out of this data. However, mining of big data comes with its own problem the major problem being the difficulty in detecting patterns in evolving data. Clustering techniques can lead us to discover hidden information in these evolving data. Traditional data clustering models such as K-means do not scale well on large data set and are dependent on assumptions regarding cluster statistical properties (ie. Number of clusters, cluster shape), while unsupervised ANN algorithms (Adaptive Resonance Theory – ART networks) are recognized widely by their ability to discover hidden patterns, adapt to evolving large data and robustness to noise. Consequently, in this paper, Adaptive Resonance Theory 2(ART2) was used to address the problem of clustering large data set using a sensor stream data for the clustering experiments. The data was categorized into 40 categories or clusters close to the 54 class of the data with initial parameters of vigilance parameter = 0.9. The vigilance parameter was varied to study the evolution of the number of categories. The pattern match (clusters) gets finer when the vigilance parameter is closer to 1 and coarser when it is away from 1. The study revealed that the closer the value of the vigilance parameter to 1, the more number of clusters (categories) produced and the farther the value of the vigilance parameter the less the number of clusters.
- ItemOpen AccessComparative analysis of text categorization algorithms(The journal of computer science and its applications, Nigeria computer society, 2017) Adewole, A.P.; Omitiran, D.M.Text categorization (also known as text classification) is the task of automatically assigning documents to a category (or categories) from a pre-specified set. This task has several applications, including spam filtering, identification of document genre, automated indexing of scientific articles according to a predefined thesauri of technical terms, and even the automated extraction of metadata. The importance of text categorization cannot be overemphasized due to the fact that unstructured texts are the largest readily available source of data and manual organization of this data is infeasible due to the large number of documents involved as well as time constraints. The accuracy of modern text categorization machines rivals that of trained human professionals. This study experimentally compared four machine learning classifiers used in text categorization. These algorithms are; Naïve Bayes, Decision trees, k-Nearest Neighbour (kNN) and Support Vector Machines (SVM). These classifiers were developed using Python programming language. When run on the Reuters dataset, SVM significantly outperforms Naïve Bayes, kNN and Decision Trees. Decision trees performed worst of the four algorithms considered in this study. From observations made during the course of running these experiments, there seems to be a trade- off between simplicity and effectiveness. In conclusion, the results of this comparative analysis prove that SVM is the most effective of the classifiers considered in this study.
- ItemOpen AccessA Comparative Study of Simulated Annealing and Genetic Algorithm for Solving the Travelling Salesman Problem(International Journal of Applied Information Systems (IJAIS), 2012-10) Otubamowo, K.; Egunjobi, T.O.; Adewole, A.P.Metaheuristic algorithms have proved to be good solvers for the traveling salesman problem (TSP). All metaheuristics usually encounter problems on which they perform poorly so the programmer must gain experience on which optimizers work well in different classes of problems. However due to the unique functionality of each type of meta-heuristic, comparison of metaheuristics is in many ways more difficult than other algorithmic comparisons. In this paper, solution to the traveling salesman problem was implemented using genetic algorithm and simulated annealing. These algorithms were compared based on performance and results using several benchmarks. It was observed that Simulated Annealing runs faster than Genetic Algorithm and runtime of Genetic Algorithm increases exponentially with number of cities. However, in terms of solution quality Genetic Algorithm is better than Simulated Annealing.
- ItemOpen AccessComparison of selected image processing techniques(Lautech Journal of engineering and Technology, 2014) Adewole, A.P.; Mustapha, M.O.Image as an important artifact faced with several constraints that may inhibit its usefulness. These constraints includes noise, Identification of objects in the image and extraction of features. In this paper, the denoising methods of Two Stages Image Denoising By Principal Component Analysis With Local Pixel Grouping(PCA - LPG) and Non Linear Filtering Algorithm For Underwater Images, the object identification methods of SCALE-INVARIANT FEATURE TRANSFORM (SIFT) and SPEEDED UP ROBUST FEATURES (SURF), the feature extraction methods of thresholding and subtraction and template matching are compared experimentally The experimental evaluation of these algorithms made it possible to draw some conclusions. These conclusions are supported from the results of the implementations of each technique, hence the recommended technique for denoising is Local Pixel Grouping (PCA - LPG), the recommended technique for object identification is SPEEDED UP ROBUST FEATURES (SURF) and the recommended technique for feature extraction is tresholding and subtraction. The recommended techniques for each of the concept were implemented in C# programming language with the help of an open source computer vision library EmguCV.
- ItemOpen AccessDigital Filter Design Using Artificial Neural Network.(Journal of Computer Science and its Applications, 2010-06) Adewole, A.P.; Agwuegbo, S.O.In this paper, Feed Forward Multi-Layer Perceptron neural network was adapted as a digital filtering tool in modelling communications systems that were corrupted by noise or interference. Discrete-Fourier Transform was used to reduce error in transmission. The input and target output data from the study were generated using ionosphere data radar, and this proved to be essential and necessary for training and testing the network. The network was trained using MATLAB R2008a and the training resulted to the minimisation of the error. The result of digit filtration shows a near error-free output. In conclusion, the forward-feed multilayered neural network can be used to build a functional digital filter.
- ItemOpen AccessDrawing Graphs with Modified Simulated Annealing Algorithm(Journal of Computer Science and its Applications, 2006-06) Adewole, A.P.This paper illustrates a modified simulated annealing algorithm for drawing graphs according to a number of aesthetic criteria. The proposed modified algorithm used in this paper combined the power of simulated annealing algorithm and mouse events available in java to enhance graph readability. Tests are carried out on two graphs of increasing difficulty, and the results show that this approach draws graph nicely and at the same time meets some of the aesthetic criteria stated in this work.
- ItemOpen AccessEvaluation Of Linear Interpolation Smoothing On Naive Bayes Spam Classifier(INTERNATIONAL JOURNAL OF TECHNOLOGY ENHANCEMENTS AND EMERGING ENGINEERING RESEARCH, 2014) Adewole, A.P.; Fakorede, O.J.; Akwuegbo, S.O.N.The inconvenience associated with spams and the cost of having an important mail misclassified as spam have made all efforts at improving spam filtering worthwhile. The Naive Bayes algorithm has been found to be successful in properly classifying mails. However, they are not perfect. Recent researches have introduced the idea of smoothing into the Naive Bayes algorithm and they have been found to produce better classification. This study applies the concept of linear interpolation smoothing to Naive Bayes spam classification. The resulting classifier did well at improving spam classification and also reducing false positives.
- ItemOpen AccessEvolutionary simulated annealing with application to tournament traveling problem(Journal of Computer Science and its Applications, 2009) Adewole, A.P.; Ejide, F.OThis paper describes evolutionary simulated annealing approach to the solution of TTP that includes both feasible and infeasible schedules. The approach uses a large neighbourhood with complex moves and includes techniques such as perturb, pervert and score t balance the exploration of the feasible and infeasible regions and to escape local minimal at very low temperatures. Experimental results show that, in practical cases, the optimal solution can be found in reasonable time, for small instances of n.
- ItemOpen AccessFORECASTING DISTRIBUTED DENIAL OF SERVICE ATTACK USING HIDDEN MARKOV MODEL(LAUTECH Journal of Engineering and Technology, 2016) Afolorunso, A.A.; Abass, O.; Longe, H. O. D.; Adewole, A.P.consumes the system resources in terms of time, memory, and processors. This paper presents a proposed method for forecasting DDoS in networks. The proposed model employs hidden Markov model (HMM) to forecast DDoS attacks. The method uses the inherent characteristic features of DDoS to determine the observable states of the system.To avoid intractable computations, Kullback-Leibler divergence algorithm was employed to reduce the number of observable states to three. The proposed model is formulated and trained through experiments using DARPA 2000 data set and the preliminary resultsshows that the characteristic features of the DDoS and the entropy concept can be used to formulate an HMM to predict DDoS.
- ItemOpen AccessHIERARCHICAL CLASSIFICATION OF MORPHOLOGICAL FEATURES OF TILAPIA CABREA(Journal of Natural Sciences, Engineering and Technology, 2012-11) Agwuegbo, S.O.N.; Adewole, A.P.; Isenah, M.G.This article proposes an effective data visualization of multidimensional data. These displays are useful to represent the existence or absence of relationships among objects corresponding to hierarchical classifications, bifurcation or evolutionary structure. The display in this article used some morphological features of Tilapia Cabrea, as represented in the dendrogram or cluster tree which illustrates the successive fusions of objects into groups or divisions made at each successive stage of the analysis. Effectively, this clustering reduces the dimensionality and makes interpretations easier.
- ItemOpen AccessA model-based collaborative filtering with dimensionality reduction(The journal of computer science and its applications, 2017-12) Adewole, A.P.; Kawedo, C.V.In this day and age, the measure of data accessible online multiplies exponentially. With such development rate, it is getting to be distinctly troublesome for clients to approach things of interest subsequently bringing about information overload issue. This overload produces information in very high dimensions and makes it challenging for these systems to suit or accommodate this increment in data. One of the issues with high-dimensional datasets is that, in many cases, not all the measured factors are "vital" for comprehending the underlying phenomena of interest. The use of mathematical procedures to tackle these problems by reducing the dimensions of the data can successfully alleviate such problems and generate more accurate recommendations. This paper proposes a Model-Based Collaborative Filtering (CF) algorithm that integrates dimensionality reduction technique to lessen known limitations of collaborative filtering techniques. The algorithm consists of building a recommender system for movies using data from the MovieLens Recommender System containing 100,000 ratings. The analytic model was constructed using the standard CRISP- DM methodology. According to the experimental results obtained, the proposed algorithm proved to be very effective as far as dealing with both the sparsity and scalability problems and thus produced more accurate predictions and recommendations when contrasted with the standard Item-based CF technique and the random CF technique.
- ItemOpen AccessNeural Network Model for Performance Evaluation of Academic Staff of Tertiary Institutions(International Journal of Applied Information Systems (IJAIS), 2013-01) Okoye-Ubaka, M.N.; Adewole, A.P.; Folorunso, O.; Ezike, J.O.J.This paper presents a neural network model for accomplishing the task of performance evaluation of academic staff of tertiary institutions. Data was collected from 10 randomly selected institutions using the completed Annual/Appraisal Performance and Evaluation Report (APER) form for academic staff. Fifteen Human Resource Metrics were considered which were classified under three main factors namely: Research, Teaching and Service. These are the major Human Resource foci of the Tertiary Institutions. The datasets were divided into three: train, validation and test data. The train data was presented to the Supervised Neural Network to approximate the fifteen Human Resource variables. The learning parameters for the training and testing of the survey data varied from 0.07 to 0.1 and the momentum parameter approached zero value (0.01 to 0.03). Root Mean Square Error (RMSE) was computed for both the parametric models (Principal Component 0.80 and Factor Analysis 0.15). The result revealed that Multilayer Perceptron Neural Network (MPNN) with back propagation algorithm got better outcome when compared with those parametric models. Experimental results in this study demonstrated that MPNN based model can closely predict the Human Resource metrics, with minimum RSME at 90%.
- ItemOpen AccessPredicting Insurance Investment: A Factor Analytic Approach(Journal of Mathematics and Statistics, 2010) Agwuegbo, S.O.; Adewole, A.P.; Maduegbuna, A.N.Problem statement: In the last decade growing attention has been paid to the pattern of investments by the insurance industry and the question of how to evaluate such investments. In an economy where the capital market is huge and active, mathematical considerations come into play in the selection of investments to ensure yield maximisation. Approach: This study examined the use of factor analysis as an emerging technique for the analysis of insurance investment in Nigeria. Results: The proposed technique described a number of methods designed to analyze interrelationships within the investment variables in terms of few underlying but unobservable random quantities called factors. The factors were constructed in a way that reduces the overall complexity of the data by taking advantage of inherent interdependencies. Conclusion: The result obtained through this approach were promising and shows that two principal components of the factor loadings have a cumulative proportion of variance accounted for 94.5% of the total variations of the investments pattern.
- ItemOpen AccessThe Quadratic Entropy Approach to Implement the Id3 Decision Tree Algorithm(Journal of Computer Science and Information Technology, 2018-12) Adewole, A.P.; Udeh, S.NDecision trees have been a useful tool in data mining for building useful intelligence in diverse areas of research to solve real world problems of data classifications. One decision tree algorithm that has been predominant for its robust use and wide acceptance has been the Iterative Dichotomiser 3 (ID3). The splitting criteria for the algorithm have been the Shannon algorithm for evaluating the entropy of the dataset. In this research work, the implementation of the ID3 algorithm using the Quadratic entropy algorithm in a bid to improve the accuracy of classification of the ID3 algorithm was carried out. The results show that the implementation of the ID3 algorithm using the quadratic entropy with some selected datasets have a significant improvement in the areas of its accuracy as compared with the traditional ID3 implementation using the Shannon entropy. The formulated model makes use of similar process of the ID3 algorithm but replaces the Shannon entropy formula with the Quadratic entropy.
- ItemOpen AccessA Random Walk Model for Stock Market Prices(Journal of Mathematics and Statistics, 2010) Agwuegbo, S.O.N.; Adewole, A.P.; Maduegbuna, A.N.Abstract: Problem statement: The stock exchange market has been one of the most popular investments in the recent past due to its high returns. The market has become an integral part of the global economy to the extent that any fluctuation in this market influences personal and corporate financial lives and the economic health of a country. The daily behavior of the market prices revealed that the future stock prices cannot be predicted based on past movements. Approach: In this study, we analyzed the behavior of daily return of Nigerian stock market prices. The sample included daily market prices of all securities listed in the Nigeria Stock Exchange (NSE). Results: The result from the study provided evidence that the Nigerian stock exchange is not efficient even in weak form and that NSE follow the random walk model. The idealized stock price in the Nigerian stock exchange is a martingale. Conclusion: Martingale defines the fairness or unfairness of the investment and no investor can alter the stock price as defined by expectation.
- ItemOpen AccessA Receipt-free Multi-Authority E-Voting System(International Journal of Computer Applications, 2011-09) Adewole, A.P.; Sodiya, A.S.; Arowolo, O.A.The existing e-voting schemes satisfied requirements such as eligibility, completeness, „no vote duplication‟, privacy but have not been able to solve the problems of universal verifiability, coercion, bribery and fairness in the overall election process. In this work, a receipt-free multi-authority e-voting system is proposed to solve the drawbacks of the existing e-voting systems is proposed. The proposed scheme employs ElGamal encryption for ensuring the security of votes because of its probabilistic nature. ElGamal which is homomorphic with multiplication is modified to be additive homomorphic in order to ensure voters‟ privacy and overall election efficiency. A trusted centre is involved in the scheme to distribute the shared secret key among the authorities and the Shamir(t, n) threshold scheme is used for key distribution. The authorities will then use this share secret to decrypt the encrypted ballot. 1-out-of-L re- encryption is used to guarantee receipt-freeness. The proposed scheme is divided into registration, validation, vote casting and tallying phases. The security analysis of the scheme was then carried out to show its effectiveness.
- ItemOpen AccessReducing the Observable States Space of Hidden Markov Model for Distributed Denial of Service Attack Prediction using Kullback-liebler Divergence.(UNILAG Journal of Medicine, Science and Technology, 2017) Afolorunso, A.A.; Adewole, A.P.; Abass, O.; Longe, H. O. D.Distributed Denial of Service (DDoS) attack floods the network with loads of unwanted packets and requests that weigh down the system resources such as memory and processors. Hidden Markov model (HMM) is one of the models that can be used to predict and detect such attacks. A problem to be solved was determining the observable states and subsequently, the model parameters since the performance of the model depends on the accurate selection of these parameters. In this work, the concept of entropy was used to determine the observable states, which characterise the HMM. In order to improve computational efficiency of the algorithm for estimating the parameters of the model, Kullback-Liebler Divergence (KLD) method was employed for reducing and selecting appropriate observable states to achieve a good prediction model. The experimental results justified the suitability of KLD in reducing the entropy-based observable states of HMM for predicting DDoS attack.
- ItemOpen AccessSignal Model for Prediction of Exchange Rates(Journal of the Nigerian Association of Mathematical Physics, 2017-05) Agwuegbo, S.O.N.; Onugha, E.E.; Akintunde, A.A.; Adewole, A.P.The appropriate prediction of exchange rates is an area of financiaL forecasting which attracts a great deal of attention. For many years, the volatile nature of exchange rates has been the focus of Many researchers. Many researchers attribute interest in exchange rate volatility to the fact that it is empirically difficult to predict future exchange rate values. Foreign exchange market generally produces observable outputs which can be characterised as signals. III this study we investigated the monthly average exchange rates of usDollars to the Nigerian Naira using signal modelling approach, Front the study, there is a convincing statistical evidence to believe that exchange rates can be better modelled by a Markov process as the output of a first order discrete autoregressive process. The result demonstrated that a Markov process is sometimes called a first order autoregressive process.
- ItemOpen AccessState-space models for analysis of hydrological series(Journal of Engineering Research, 2010-12) Agwuegbo, S.O.N.; Adewole, A.P.In this study, the class of state pace model for which optimal forecasts may be computed using a recursive estimation procedure called the Kalman Filter is developed for the analysis of hydrological series. The state-space formulation yields a practical means of estimation for his complex time varying dynamical process. It provided a generic and flexible framework for hydrological modelling and inference. A straight forward implementation was achieved in the software package S-Plus
- ItemOpen AccessStructural Model for the Analysis of Stock Market Price Index(International Journal of Marketing Studies, 2011) Agwuegbo, S.O.N.; Mojekwu, J.N.; Adewole, A.P.; Maduegbuna, A.N.Stock market returns are predictable from a variety of financial and macroeconomic variables and have long been an attraction for equity investors. Recently increasing attention has shifted on the market index as a method of measuring a section of the stock market. The stock market index is regarded as an important indicator by the investing public at large and can be used as a benchmark by which investor or fund manager compares the returns of their own portfolio. In this study, attempts are made to model the Nigerian stock market index using a structural model. The procedure is based on the relationship between the state space and the autoregressive moving average (ARMA) model. The time series procedure from S-PLUS software is used in the analysis. The result obtained shows that the Nigerian stock market price index is an autoregressive model (AR) of order 1. It is also found that the AIC is at minimum at lag 1which corresponds to the same model identified for the series by using the sample ACF and PACF.