Weather Pattern Extraction using Statistical Methods & Machine Learning Techniques

Abstract

It is known to all that the existence of life is highly dependent on the weather. Due to the unfavorable condition of current global weather the existence of life is in danger already. Since for the existence of lives a livable environment which lies in the weather is very much intrinsic, it should be taken care of before it is too late. This is the main context of this research. The goal of this research is to find out the future condition of temperature of some particular places of California using machine learning and statistical methods and compare which place will be more livable after two years. Currently, one of the most alarming issue in the world is the global warming. The effect of global warming is increasing rapidly every day without any sign of slowing down. As a result of this, it’s very concerning and important to understand the state of the temperature of the world and the route it will take in the future. As such, the objective of this reseach is to predict the temperature conditions of the future. The research starts by collecting data of few select areas in california and hence, extracted data from 14 stations of california. The data was then fed to the ARIMA model to find the future trend with the respective ARIMA orders and other paremeters per station. The research has successfully identified the trend of the next 730 days (2 years) while considering the errors that the model creates. Furthermore, the research tried to identify the most favorable place to live, in california, by comparing the RMSE of the different stations by comparing the distance between the favorable human ambient temperature of 70◦F with the results that we got from the prediction. As such, the ‘Miramar’ station gave the least RMSE value of 10.7824 while the ‘lake Arrowhead’ gave the worst RMSE of 24.3605. From these RMSE values and also the learning curves it was decided, the most favorable place to live around was the ‘Miramar’ station, while ‘lake Arrowhead’ station was the worst in terms of favourable temperature for humans.