Mohammed Joda Usman

Orthogonal Wavelet Support Vector Machine for Predicting Crude Oil Prices

Previous studies mainly used radial basis, sigmoid, polynomial, linear, and hyperbolic functions as the kernel function for computation in the neurons of conventional support vector machine (CSVM) whereas orthogonal wavelet requires less number of iterations to converge than these listed kernel functions. We proposed an orthogonal wavelet support vector machine (OSVM) model for predicting the monthly prices of West Texas Intermediate crude oil prices. For evaluation purposes, we compared the performance of our results with that of the CSVM, and multilayer perceptron neural network (MLPNN). It was found to perform better than the CSVM, and the MLPNN. Moreover, the number of iterations, and time computational complexity of the OSVM model is less than that of CSVM, and MLPNN. Experimental results suggest that the OSVM is effective, robust, and can efficiently be used for crude oil price prediction. Our proposal has the potentials of advancing the prediction accuracy of crude oil prices, which makes it suitable for building intelligent decision support systems.

Co — Active neuro-fuzzy inference systems model for predicting crude oil price based on OECD inventories

This paper present a novel approach to crude oil price prediction based on co-active neuro-fuzzy inference systems (CANFIS) instead of the commonly use fuzzy neural network and adaptive network-based fuzzy inference systems due to superiority and robustness of the CANFIS model. Monthly data of West Texas Intermediate crude oil price and organization for economic co-operation and development (OECD) inventories, obtained from US Department of Energy were used to built the propose model. The CANFIS prediction model was trained, validated and tested. The performance of our approach is measured using mean square error, root mean square error, mean absolute error and regression. Suggestion from the results shows that the CANFIS demonstrated a high level of generalization capability with relatively very low error and high correlation which exhibited successful prediction performance of the proposal. The model has the potential of being developed into real life systems for use by both government and private businesses for making strategic planning that can boost economic activities.

Optimization of neural network through genetic algorithm searches for the prediction of international crude oil price based on energy products prices

This study investigated the prediction of crude oil price based on energy product prices using genetically optimized Neural Network (GANN). It was found from experimental evidence that the international crude oil price can be predicted based on energy product prices. The comparison of the prediction performance accuracy of the propose GANN with Support Vector Machine (SVM), Vector Autoregression (VAR), and Feed Forward NN (FFNN) suggested that the propose GANN was more accurate than the SVM, VAR, and FFNN in the prediction accuracy and time computational complexity. The propose GANN was able to improve the performance accuracy of the comparison algorithms. Our approach can easily be modified for the prediction of similar commodities.

Energy-Efficient Resource Allocation Technique Using Flower Pollination Algorithm for Cloud Datacenters

Cloud Computing is modernizing how Computing resources are created and disbursed over the Internet on a model of pay-per-use basis. The wider acceptance of Cloud Computing give rise to the formation of datacenters. Presently these datacenters consumed a lot of energy due to high demand of resources by users and inefficient resource allocation technique. Therefore, resource allocation technique that is energy-efficient are needed to minimize datacenters energy consumption. This paper proposes Energy-Efficient Flower Pollination Algorithm (EE-FPA) for optimal resource allocation of datacenter Virtual Machines (VMs) and also resource under-utilization. We presented the system framework that supports allocation of multiple VMs instances on a Physical Machine (PM) known as a server which has the potential to increase the energy efficiency as well resource utilization in Cloud datacenter. The proposed technique uses Processor, Storage and Memory as major resource component of PM to allocate a set of VMs, such that the capacity of PM will satisfy the resource requirement of all VMs operating on it. The experiment was conducted on Multi-RecCloudSim using Planet workload. The results indicate that the proposed technique energy consumption outperform the benchmarking techniques which include GAPA, and OEMACS with 91% and 94.5% energy consumption while EE-FPA is around 65%. On average 35% of energy has been saved using EE-FPA and resource utilization has been improved.