Abstract:
As global urbanization and climate change accelerate, Dengue fever is spreading globally.
Bangladesh has also experienced varying degrees of Dengue fever, particularly in the city of
Dhaka, causing huge economic losses. Therefore, we collected data on temperature, relative
humidity, and rainfall in Dhaka city and tried to find out what kind of relationship there is with
dengue. For this purpose, we have collected data on the accuracy of dengue fever forecast in Dhaka
city during the period 2010-2019 and also collected our weather data for the same period. First,
apply the Linear Regression algorithm of machine learning to this data set to investigate the
association of climate with dengue fever. Then apply the Time Series algorithm to whom I have
tried to clarify how it is influencing over time. So in our work, we first use the time series dengue
fever data that were decomposed into seasonal, trend, and remainder components. Now the
seasonal-trend decomposition procedure is based on loess (STL). Then secondly, the time lag of
variables was determined in cross-correlation analysis and the order of autocorrelation was
estimated using autocorrelation (ACF) and partial autocorrelation functions (PACF). Finally, the
two algorithms performed very well on our datasets. Applying the time series algorithm was very
challenging for us because we know that Dengue fever is mainly in August, September, and
October of the year. September, and October is the maximum but at other times their effect is less.
Also we convert our data into categorical and apply some other algorithms. One of them is
Logistics Regression, Decision Tree, Navie Bayes, K-Nearest Neighbors (KNN), Support Vector
Machine (SVM) and also apply Random Forest Regression. But here some Algorithm works well
but the result of some Algorithm was not satisfactory. Besides that, the biggest challenge was data
collection. But in the end, i succeeded and fully did it.
