Search

The rule-based approach for the analysis and interpretation of stroke symptoms for the purpose of accurate and timely diagnosis and treatment has been reported as weak. In this paper, an artificial neural network using a sigmoid activation function is used to develop a stroke diagnosis system that is capable of automatically extracting the features responsible for stroke. The system is capable of using knowledge learned from data to predict the likelihood of a patient having a stroke or not. The efficiency of the stroke prediction system was evaluated on live data obtained from four medical centers located in the southwestern part of Nigeria. The results show an average prediction accuracy of 93.75% and can carry out a complete diagnosis of a patient within 4 microseconds. The proposed system, therefore, will facilitate quick diagnosis and timely treatment of stroke patients and consequently has the potential of reducing brain damage and other complications that may arise due to slow or delayed diagnosis of strokes. [ABSTRACT FROM AUTHOR]

The difficult jobs of investigating illegal activity and identifying offenses have fallen to the Law Enforcement Agencies (LEA). To help the LEA solve crimes, a number of criminal profiling systems have been created; however, the methods used in most systems do not allow for the clustering of criminals according to their behavioral traits. In order to get the best SVM classifier with the best kernel function that best fits our crime data set, this study hybridized the Fuzzy C-Means technique with the support vector machine algorithm in criminal profiling. To reduce the intra-cluster variances, the Fuzzy C-Means (FCM) method was altered by hybridizing it with Support Vector Machine (SVM). This was accomplished by substituting the SVM inner-product distance norm for the Euclidean distance used in the current FCM method to calculate the similarity and dissimilarity measure. Next, additional characteristics were added to the data along with the hybridized algorithm. MATLAB scripts were used to implement the developed strategies. The following criteria assessed the performance: execution time, sensitivity, precision, accuracy, and specificity. The result shows that the RBF kernel function performed best with both OAA and OAO classifiers. The OAA classifier performed best with 93.53% Specificity, 96.89% Precision, and 95.44% Accuracy over OAO and the pairwise classifier (BSVM). Therefore, the RBF kernel function using the OAA classifier is recommended to best suit our crime data set for criminal profiling, contributing to Sustainable Development Goal (SDG) 16. [ABSTRACT FROM AUTHOR]

The performance of two classification techniques, logistic regression and Support Vector Machines (SVMs), in assessing vaccination data is investigated in this study. The model was trained based on leave-out-one cross validation to obtain an accurate result. Simulated with ten thousand replications, a life data set was used to establish a better model. The findings from the simulation revealed that the logistic regression model slightly outperformed the SVM while the life data shows that the tuned SVM outperformed both the logistic and the SVM. This demonstrates the practical utility of advanced approaches such as SVMs in difficult categorization scenarios such as vaccination prediction. The study emphasizes the superiority of the customized SVM model in this setting, as well as the potential of machine learning approaches to increase comprehension of complicated healthcare scenarios and guide data-driven decision-making for influencing vaccination plans and public health. The study recommends the use of logistic regression if the data point is high. [ABSTRACT FROM AUTHOR]

The recent evolution of the novel COVID-19 pandemic posed unprecedented challenges for the global community and researchers alike without reference literature or prior prediction approaches applied to that context, leading to its great impact on the economies of different countries. This work aims to address these challenges by predicting the spread of COVID-19 in Nigeria. In order to carry out this research, raw data on COVID-19 pandemic in Nigeria was gotten from John Hopkins University repository through Github as submitted by the Nigeria Center for Disease Control (NCDC) through the World Health Organization (WHO). Python, along with its packages such as Numpy, SciPy, Pandas, Matplotlib, and Plotly, was used to process and visualize the data for this model. Google Colab; a Jupyter notebook environment that run on the cloud was employed too. The SIR (Susceptible, Infected, Recovered or Removed) population model, which is commonly employed for modelling epidemics, was used to predict the spread of the pandemic in Nigeria. The machine learning Python packages used in the analysis included Pandas for data analysis and manipulation, Numpy for executing mathematical and logical operations, and SciPy for operations such as regression, minimization, and Fourier series. The model developed in this study predicted a peak of the pandemic between June and August, 2020, and a flattening of the curve in January 2021; the predicted dates were accurate when compared to real-world data. The outcome of the study provides better insight (for instance; rate of spread in response to certain conditions and corresponding mortality or survival rate) on the future of the COVID-19 pandemic; yielding more information that will help in planning for the future through evidenced correctness and reliability. The work present the recovery coefficient rate (β), the transmission coefficient (γ) and, the basic reproduction ratio (R0) values for the data under consideration that indicates a reliable model for predictions. [ABSTRACT FROM AUTHOR]

In this work, a multiple dependent state sampling plan, which is an inspection procedure that determines whether an attribute is conforming or non-conforming to a specific requirement, in which the decision criterion for each lot dictates whether to accept the lot; reject the lot; or conditionally accept or reject the lot based on the disposition of future related lots, is introduced. This plan has some advantages over other acceptance sampling plans, like increased efficiency, improved ability to discriminate between acceptable and non - acceptable lots or batches, flexibility in designing the sampling process, and improved cost-effectiveness. To reject a lot, the plan made use of the properties of the sampled current and preceding lots. The study aims to reduce the average sampling number by using a non-linear optimization problem that is subjected to some constraints. With regards to a life test that is truncated in time, the product's median life was used for the proposed sampling plan assuming that the lifetime of the product follows Zech distribution. The usage of median life was necessitated because Zech distribution is an asymmetric distribution with longer tail to the right. Two points on the operating characteristic curve were used for the proposed sampling plan and the following parameters were found; number of preceding lots which is required for deciding if the current lot should be accepted or rejected, the size of the sample, rejection number, and acceptance number. For different shape parameters, we constructed tables for various combinations of consumers' and producers' risks. A real example was provided which showed that a multiple dependent state sampling plan is a good sampling plan for fitting the datasets. Comparing the proposed plan with a single sampling plan, the results reveal that the proposed plan is more effective at securing the consumer and the producer with less inspection. The approach introduced in this study provides an ample opportunity for the manufacturers to reduce the cost and time of inspection by having the sample size reduced without compromising the decision-making accuracy. By implementing the findings of this study, the consumers are confident that their hard-earned money is not used to purchase sub-standard goods. [ABSTRACT FROM AUTHOR]

This paper is used to propose a reparametrised autoregressive model that is capable of analyzing time series data that follows a non-Gaussian marginal distribution. The Anderson Darling Statistics was used to identify that Nigerian Gross domestic product followed a Gamma distribution. The proposed Gamma autoregressive (GAR) and classical autoregressive models were fitted using a Maximum Likelihood Estimation (MLE) method. The Akaike Information Criteria (AIC) was used to select AR(2) and GAR(2) as the optimal models but GAR(2) was chosen because it has the least value of AIC. The comparison between AR(2) and GAR(2) models based on the values of Mean absolute error (MAE), Mean absolute prediction error (MAPE) and Root mean square error (RMSE) indicted that GAR(2) will yield a more accurate forecast than AR(2). In essence, GAR model is a viable alternative and better model for analyzing GDP growth rate. [ABSTRACT FROM AUTHOR]

Most frequently used models for modeling and forecasting periodic climatic time series do not have the capability of handling periodic variability that characterizes it. In this paper, the Fourier Autoregressive model with abilities to analyze periodic variability is implemented. From the results, FAR(1), FAR(2) and FAR(2) models were chosen based on Periodic Autocorrelation function (PeACF) and Periodic Partial Autocorrelation function (PePACF). The coefficients of the tentative model were estimated using a Discrete Fourier transform estimation method. FAR(1) models were chosen as the optimal model based on the smallest values of Periodic Akaike (PAIC) and Bayesian Information criteria (PBIC). The residual of the fitted models was diagnosed to be white noise. The in-sample forecast showed a close reflection of the original rainfall series while the out-sample forecast exhibited a continuous periodic forecast from January 2019 to December 2020 with relatively small values of Periodic Root Mean Square Error (PRMSE), Periodic Mean Absolute Error (PMAE) and Periodic Mean Absolute Percentage Error (PMAPE). The comparison of FAR(1) model forecast with AR(3), ARMA(2,1), ARIMA(2,1,1) and SARIMA(1,1,1 )(1,1,1)12 model forecast indicated that FAR(1) outperformed the other models as it exhibited a continuous periodic forecast. The continuous monthly periodic rainfall forecast indicated that there will be rapid climate change in Nigeria in the coming yearly and Nigerian Government needs to put in place plans to curtail its effects. [ABSTRACT FROM AUTHOR]