-
Modeling the Impact of Climate Factors on COVID-19 Transmission in Nigeria
Audu Musa Mabu,
Babagana Modu,
Babagana Ibrahim Bukar,
Musa Ibrahim Dagona
Issue:
Volume 8, Issue 2, April 2022
Pages:
23-29
Received:
22 January 2022
Accepted:
4 March 2022
Published:
15 March 2022
Abstract: Due to spatial and temporal changes in climate, the incidences of COVID-19 is much more higher in some parts of America, Europe and Asia by comparing with Saharan and sub-Saharan Africa. Several studies show the link between climate factors (e.g., temperature rainfall and humidity) and COVID-19 occurrence will be used to aid intervention planning, prevention and control policies. Nigeria is a country that is sensitive to spatial and temporal variability in the occurrence of climate factors, and fully knowing it link with COVID-19 is crucial towards mitigation. In this study, we examined the link by firstly deployed convenience sampling to select three cities (Abuja, Kano and Lagos) where the international airports of Nigeria are situated and also the index case of the country came through Lagos. Secondly, we used the reported cases of COVID-19 from its onset in the country (22/02/2020) up to 21/05/2021. Thirdly, lagged regression was used to explore the link between weekly counts of COVID-19 cases and weekly recorded average of the climate data; including the google trend index as a measure of the populace health seeking behaviour. We found a significant influence of temperature, humidity and heath seeking trend, with a very negligible contributions of precipitation to the occurrence of the COVID-19 in the states investigated. This result will assist policy makers with a prior knowledge to plan for non-pharmaceutical interventions in anticipation of possible outbreak.
Abstract: Due to spatial and temporal changes in climate, the incidences of COVID-19 is much more higher in some parts of America, Europe and Asia by comparing with Saharan and sub-Saharan Africa. Several studies show the link between climate factors (e.g., temperature rainfall and humidity) and COVID-19 occurrence will be used to aid intervention planning, ...
Show More
-
An Application of Multi-label Linear Discriminant Analysis and Binary Relevance K-Nearest Neighbor in Multi-label Classification of Annotated Images
Festus Malombe Mwinzi,
Thomas Mageto,
Victor Muthama
Issue:
Volume 8, Issue 2, April 2022
Pages:
30-37
Received:
4 March 2022
Accepted:
24 March 2022
Published:
31 March 2022
Abstract: Although Binary Relevance (BR) is an adaptive and conceptually simple multi-label learning technique, its inability to exploit label dependencies and other inherent problems in multi-label examples makes it difficult to generalize well in the classification of real-world multi-label examples like annotated images. Thus, to strengthen the generalization ability of Binary Relevance, this study used Multi-label Linear Discriminant Analysis (MLDA) as a preprocessing technique to take care of the label dependencies, the curse of dimensionality, and label over counting inherent in multi-labeled images. After that, Binary Relevance with K Nearest Neighbor as the base learner was fitted and its classification performance was evaluated on randomly selected 1000 images with a label cardinality of 2.149 of the five most frequent categories, namely; "person", "chair", "bottle", "dining table" and "cup" in the Microsoft Common Objects in Context 2017 (MS COCO 2017) dataset. Experimental results showed that micro averages of precision, recall, and f1-score of Multi-label Linear Discriminant Analysis followed by Binary Relevance K Nearest Neighbor (MLDA-BRKNN) achieved a more than 30% improvement in classification of the 1000 annotated images in the dataset when compared with the micro averages of precision, recall, and f1-score of Binary Relevance K Nearest Neighbor (BRKNN), which was used as the reference classifier method in this study.
Abstract: Although Binary Relevance (BR) is an adaptive and conceptually simple multi-label learning technique, its inability to exploit label dependencies and other inherent problems in multi-label examples makes it difficult to generalize well in the classification of real-world multi-label examples like annotated images. Thus, to strengthen the generaliza...
Show More
-
Deep Convolution Neural Network with Dropout in Modeling Exchange Rate Volatility
Samuel Wanjiru,
Anthony Waititu,
Anthony Wanjoya
Issue:
Volume 8, Issue 2, April 2022
Pages:
38-46
Received:
5 March 2022
Accepted:
24 March 2022
Published:
31 March 2022
Abstract: Exchange rate data possesses time-series features such as a trend. Based on a convolutional neural network (CNN) deep learning algorithm, which has the advantages of detecting patterns, extracting effective features, finding interdependence of time series data, and its computational efficiency, this paper proposes a convolutional neural network with dropout model-based approach to model and forecast exchange rates. In the meantime, this paper uses the CNN to first model and predict exchange rates and the corresponding results of this model are compared with those of the CNN-WD. The experimental results showed that the CNN-WD is superior to the CNN model in terms of the error value, fitting degree and training time. The dataset used for this research is that of daily exchange rates for the period between December 1, 2003, and October 15, 2021, which is comprised of 6528 daily trading observations. Adjusted closing rates are chosen. First, this paper adopts a CNN to effectively identify patterns and extract relevant data features of the exchange rate dataset, making use of the past 21 days. Dropout regularization is then adopted to help prevent the CNN model from overfitting data by temporarily removing a neuron from the network along with all its incoming and outgoing connections during training if its generated random value is below the set dropout rate. This paper further evaluates the reducibility and identifiability of the CNN-WD. As an application, this paper uses the CNN-WD to forecast the next month’s average tea price in Kenya.
Abstract: Exchange rate data possesses time-series features such as a trend. Based on a convolutional neural network (CNN) deep learning algorithm, which has the advantages of detecting patterns, extracting effective features, finding interdependence of time series data, and its computational efficiency, this paper proposes a convolutional neural network wit...
Show More
-
A Hybrid Classification Model of Artificial Neural Network and Non Linear Kernel Support Vector Machine
Lena Anyango Onyango,
Anthony Gichuhi Waititu,
Thomas Mageto,
Mutua Kilai
Issue:
Volume 8, Issue 2, April 2022
Pages:
47-58
Received:
8 March 2022
Accepted:
28 March 2022
Published:
22 April 2022
Abstract: Machine Learning Algorithms are employed in characterization, pattern recognition, and prediction. A hybrid model helps in reducing the computational complexity, improves accuracy, and results in an effective method for classification. The misclassification of the individual classifier is often excluded in a hybrid classifier. The objective of this research was to develop a hybrid classification model of Artificial Neural Network and non-linear kernel Support Vector Machine as an intelligent tool for achieving better classification performance and minimizing error rates. This study further evaluated the irreducibility and identifiability statistical properties of the ANN-SVM model. To achieve the hybridization of ANN and SVM, the research first obtained weights from the fitted Support Vector Machine model, and these weights were used as the initial weights in the Artificial Neural Network structure. The experiment was carried out in three distinct phases: selection of input features using the Boruta Wrapper Algorithm, classifier learning, and classifier combined effect and classification optimization. The study findings suggest that the hybrid ANN-SVM approach gives a higher performance accuracy of 89.7% and is more precise as compared to single ANN, SVM data mining algorithms. Therefore, the hybrid of ANN-SVM is the best binary classification system for classifying diabetes mellitus. The statistical software used for analysis was R.
Abstract: Machine Learning Algorithms are employed in characterization, pattern recognition, and prediction. A hybrid model helps in reducing the computational complexity, improves accuracy, and results in an effective method for classification. The misclassification of the individual classifier is often excluded in a hybrid classifier. The objective of this...
Show More
-
Prediction of Levels of Happiness Using Multinomial Logit as a Neural Network: Evidence in Kenyan Youths
Martin Kinyua Ngari,
Anthony Kibera Wanjoya,
John Mwaniki Kihoro
Issue:
Volume 8, Issue 2, April 2022
Pages:
59-71
Received:
22 March 2022
Accepted:
9 April 2022
Published:
26 April 2022
Abstract: Happiness has become a major concern across many disciplines starting form public policy, economics and psychology because of the effects that come with not being happy. Psychologist would want to know the effects of low levels of happiness, economist would want to know the effects of levels of happiness in to the market place, researchers from health would be concerned with effects of high and low levels of happiness to health status. While predominantly, people had just a philosophical notion about happiness, currently there are numerous scientific studies on happiness. Approaches like cluster analysis have been employed before. This research used neural networks to classify multinomial levels of happiness of Kenyan youths by considering life style aspects of current life such as Internet usage, Physical activeness, Health, Social life, Education, Income, Country’s top leadership, Dining and Sleeping Habits. The research was able to fit a 14-1-4 neural network model to classify levels of happiness in Kenyan youths, an accuracy of 73% was achieved. The data was randomly split in to 70% training set and 30% test set. The training set was balanced using SMOTE approach. This research trained the model by applying gradient descent using error back propagation algorithm with initial weights drawn from uniform distribution and applied softmax activation function. Accuracy metrics were confusion matrix, precision and recall for each level of happiness, and F-Scores. The top leading factor related to happiness positively was physical activeness with youths who were more active being happier. The second factor was relationship type, the married youths were happier than the singles, separated or engaged. Youths who were more satisfied with their relationship, they were happier. Health was also positively related to happiness. On the other hand, the number of hours a youth spent on social media negatively affected their levels of happiness. The more the number of hours the low levels of happiness.
Abstract: Happiness has become a major concern across many disciplines starting form public policy, economics and psychology because of the effects that come with not being happy. Psychologist would want to know the effects of low levels of happiness, economist would want to know the effects of levels of happiness in to the market place, researchers from hea...
Show More