An Ensemble Approach to Predict Weather Forecast using Machine Learning

Weather changes have an incredibly negative impact on the environment and triggers natural disasters all of a sudden. To forecast these changes, there are several machine learning techniques and algorithms through which the weather changes can be predicted earlier. It has been noted that, from previous analysis there are many other approaches available for weather prediction. Based on those, various parameters like temperature, humidity, wind direction, precipitation, evaporation etc are considered. After surveying the emerging techniques and datasets, a proposed system is inculcated to include the approaches such as linear regression, bayes classifier, support vector machine and decision trees. In this the bagging, boosting, decision tree, random forest and stacking algorithms are used to predict the efficient accuracy. Bagging and boosting algorithms use same base learners whereas stacking uses different base learners. The learning capacity of stacking algorithm is different so that each individual learner can learn differently about various parameters and accuracy will increase when compared with other ensemble methods. Through the study it has been concluded to implement a proactive disaster recognition system to avoid the future loss of human lives and related environmental effect.