No abstract

Machine Learning (ML) techniques for time series prediction are becoming increasingly accurate and helpful, particularly in considering climate change. As more methods are developed, it follows that differentiating between them is becoming increasingly more important as well. This work took a local temperature time series as a dependent variable and a collection of relevant climatology time series as independent variables and applied leading Machine Learning methods to them. The six methods tested included four simple models: Linear Regression (L.R.), k-Nearest Neighbor (kNN), Support Vector Machine (SVM), and Artificial Neural Network (ANN), In addition of two ensemble model methods: Random Forest (R.F.) and Adaptive Boosting (AdB). Results compared all the method's training and predictive performances to evaluate the method's overall effectiveness in forecasting the average daily temperature value. Actual data was used to train each of the mentioned ML methods, and then they were used to predict the future temperature in the study area. The analysis revealed that out of the six methods tested, the Artificial Neural Network outperformed the others in both training and prediction of temperature values in the Memphis, TN climate.

No abstract

No abstract

No abstract