click here to download project abstract
ABSTRACT
Hurricanes are powerful and deep tropical cyclones that came along could cause
serious human and property losses. The causes of hurricane are complex and
factors affecting hurricane intensity vary. The current intensity predictions are based
on hurricane movement patterns and statistical methods are greatly influenced by
expert experience, so prediction results are not satisfying. Although machine
learning and statistical methods have been extensively used to study cyclones. The
statistical analysis of tropical cyclone is very crucial. The current dynamic statistical
analysis of tropical cyclone could computationally blow up and have more
enhancement on the prediction. This research is based on ensemble learning for
predicting tropical cyclone. A multi-dimensional element choice layer is proposed in
our technique to pick the most associated meteorological factors and time reach to
typhoon tracks from the point of view of property relationship examination and
transient connection investigation. This layer is fundamental for assessing the
effects of meteorological factors on typhoon track determining.
