Oyedeji Osowole

A model for estimating the total prison population, Pt , in Oyo State, Nigeria is proposed and a probable way of projecting for future years using an appropriate ARMA model is discussed. By using a monthly dataset for four consecutive years we observed that there is a decline in prison population, perhaps due to the impact of current crime reduction strategies. The prisoners' average serving time varies with gender between 0.7-1.3years.

Poverty analysis has relied heavily on data in summarized form and this has created dearth of knowledge on the statistical properties of Foster-Greer-Thorbecke (FGT) poverty indices. This study derived estimators of FGT poverty indices from first principles in an attempt to provide an insight into some intrinsic characteristics of FGT indices. The estimators are found to be reasonably unbiased and consistent. The estimates of the indices obtained from the estimators are approximately 53%, 22% and 12 % for the head count, poverty gap and square poverty gap indices. From the conventional method, the estimates are approximately 52%, 21 % and 11% respectively. The results therefore establish the validity of the derived estimators as adequate alternative measures of the three basic poverty dimensions of proportion, depth and severity.

This study applied the approach of Response Surface Methodology to optimize some selected process factors in wheat production in order to minimize the amount of wheat harvest loss using combine harvester. The factors considered were Grain Moisture Content (GMC), Rotor Speed (RS) and Ground Speed (GS) of the combine harvester. The amount of wheat loss was evaluated in different levels of GMC (10, 15 and 20 %), RS (450, 550 and 650 rpm) and GS (3.5, 4.5 and 5.5 km/h) and the relationship between the specified factors and the amount of wheat harvest loss was established. The study revealed that the 2-factor interaction model with coefficient of multiple determination (R^2) and adjusted coefficient of multiple determination (R ̅^2) of 79.5% and 76.88% respectively and an insignificant lack of fit (p value of 0.289) best predicts the wheat harvest loss using combine harvester. In the canonical analysis, the Eigenvalues were 10.83, 4.46 and 1.21 which shows that the points 15% for grain m...

DOI: http://dx.doi.org/10.24018/ejers.2021.6.1.2028 Vol 6 | Issue 1 | January 2021 119 Abstract — This study considered a newly proposed Exponentiated Complementary Mukherjee-Islam distribution obtained by exponentiating the Complementary MukherjeeIslam distribution. Some properties of the new distribution were derived and results from the distribution indicated that the distribution is a better alternative than its baseline distribution. The new distribution is therefore a creditable addition to the existing family of exponentiated distributions.

The misuse of toxic fungicides by indigenous cocoa farmers in Nigeria stem from their inability to predict the time for black pod disease (BPD) outbreak. Prediction of possible time for BPD outbreak will provide spotlight on areas under massive BPD invasion, minimise fungicide misuse and increase control accuracy. The Multiple Regression Model (MRM): Y=α+β1X1+β2X2+…+βnXn where Y is Nx1 matrix of response variable, X1,X2,…Xn are NxK matrices of regressors, and β1,β2,…βn regression coefficients was used in model development. Eight models (MRM1-MRM8) were fitted from real life BPD data. The performances of the models were ascertained using SER, RMSEpred and R-SqAdj. Prediction(s) made by the best fitted model was compared to real life observations (Monthly BPD Occurrence (MBO), Total Annual Occurrence (TAO), and Average Annual Occurrence (AAO), respectively). The preferred model was MRM5 (ETAPOD) followed by MRM4, MRM1, MRM2, and MRM3 in terms of SER (0.22, 0.39, 0.45, 0.45 and 0.45), ...

This paper reports on application of bootstrap nonlinear regression method to a design of an experiment dataset with fewer experimental runs. Design with desired properties was augmented and verified using graphical techniques. The augmented design with the desired properties benefited the accuracy of the approximated function used. The computation power of R-language and SAS for computing nonlinear function and bootstrap was also compared.

The long-time debate between Bayesians and Frequentists is unusual among philosophical arguments. This paper used a Monte Carlo approach to evaluate the performance of Bayesian and Frequentist estimators of a binomial proportion. The result showed that the Bayesian estimator outperformed the Frequentist estimator even when judged by the Frequentist criteria of mean squared error.