Dengue fever prediction and analysis
Dengue fever is a mosquito-borne disease usually found in tropical areas such as Africa, South American and Asia, and the current annual global estimate for dengue-related cases has risen from $8.9 billion and will only keep rising from here. Given the lack of information and a predictive model for finding out the number of cases within a specific geographical location. This paper aims to develop a model for predicting dengue fever based on the climate conditions of a specific location. This is done using dengue fever data and climate data from Peru to train and test our prediction model. We perform exploratory data analysis on our model to see the relationships between the various variables and how they affect prediction capabilities. This paper explores the impact of climate change and vegetation indices on the spread of dengue. Analysis on the climate variables such as temperature, dew point, precipitation and rainfall as well as vegetation indices from the north, south, east and west of both San Juan and Iquitos. The project uses time series plot to predict the number of future cases and uses a series of descriptive plots to explain the correlation between climate, vegetation and the number of cases. The final model with the best error rate had a Mean Absolute Error of ).071 and was the combination of Principal Component Analysis and the Light Gradient Boosting model. The final analysis showed that only climate variables had an impact on the prediction of the disease.
Undergraduate thesis submitted to the Department of Computer Science and Information Systems, Ashesi University, in partial fulfillment of Bachelor of Science degree in Management Information Systems, May 2022