Comparison of Regression Algorithms to Predict Average Air Temperature

Abstract

Regression algorithms are statistical techniques used to predict the value of a dependent variable, based on one or more independent variables. These algorithms are commonly used in fields such as economics, finance, and engineering. Temperature prediction is a specific application of regression analysis. In this case, the dependent variable is temperature and the independent variables include factors such as humidity, speed of the wind, direction of the wind, and precipitation. There are many different types of regression algorithms, each with its strengths and weaknesses. The study compares the performance of multiple regression models in predicting the average air temperature, using one month's weather data for the Beşiktaş district of Istanbul. A total of 6 different regression models, including ridge, lasso, linear, polynomial, random forest (RF), and support vector (SV) regressions, were included in the study. Among the regression models trained and tested on two different data sets, the three most successful models in predicting average air temperature were lasso, RF, and polynomial regressions (PRs), respectively.

Analysis of Machine Learning Algorithms for Prediction of Short-Term Rainfall Amounts Using Uganda’s Lake Victoria Basin Weather Dataset

“…Random forest regression is an ensemble supervised learning model that predicts by selecting individual multiple decision tree regressors at random [33]. The number of trees, input variables, and node size all have an impact on the RF regression model's efficacy.…”

“…It helps with handling complex interactions since it enables the SVR model to adjust flexibly to complex patterns in the data. To make sure the SVR model performs well when applied to previously unseen data, it is essential to cross-validate the model's results using these hyperparameters on a different validation set [33]. Table 8 provides a detailed information on the tuned parameters for SVR model.…”