A Framework for Predictive Analysisof Malaria Dispersion
| dc.contributor.advisor | Atnafu, Solomon (PhD) | |
| dc.contributor.author | Kassahun, Abreham | |
| dc.date.accessioned | 2019-12-13T06:59:37Z | |
| dc.date.accessioned | 2023-11-04T12:22:49Z | |
| dc.date.available | 2019-12-13T06:59:37Z | |
| dc.date.available | 2023-11-04T12:22:49Z | |
| dc.date.issued | 6/3/2018 | |
| dc.description.abstract | As most researches have shown that health related information is inaccessible in developing counties like Ethiopia, and when it is accessible it is not used for decision making. Analysing health information for decision making has an immense contribution in combating societal health problems especially diseases like Malaria which is deadly in developing countries. Qualitative and quantitative researches have been made so far on malaria dispersion to model prediction of malaria dispersion based on various determinants. Invariably, the models considers climate as a factor of malaria dispersion but other variables such as population density, surface hydrology, living standard, population income, GDP, life span of mosquito are considered unevenly. In this work, the original datasets of malaria dispersion, climate, NDVI and population from year 2009 to 2015 are transformed into a dataset of 3671 cases so as to train algorithms to produce a model that helps to predict malaria dispersion. Among the trained algorithms, multilayer perceptron using softplus as activation function is selected as a best algorithm with a correlation coefficient of 0.9503 by using the evaluation outcome of 10 fold cross validation. The prototype implementation of the system that uses the prediction algorithm has a design pattern to facilitate the construction of an extensible system in order to integrate new algorithms into a system in such a way that modelers can use the system to upload their algorithm and build a specific model by using the existing data. Once the models are built, the system has a capability for users to select a model and to provide inputs so that the system will provide the malaria dispersion result along with the evaluation outcome. | en_US |
| dc.identifier.uri | http://etd.aau.edu.et/handle/123456789/20352 | |
| dc.language.iso | en | en_US |
| dc.publisher | Addis Ababa University | en_US |
| dc.subject | Malaria | en_US |
| dc.subject | Malaria Dispersion | en_US |
| dc.subject | Prediction Model | en_US |
| dc.subject | Predictive Analysis | en_US |
| dc.subject | Multilayer Perceptron | en_US |
| dc.title | A Framework for Predictive Analysisof Malaria Dispersion | en_US |
| dc.type | Thesis | en_US |