Preprint / Version 1

Comparative Analysis and the Progression of Different Machine Learning Models in the field of Meteorology


  • Sriram Yerramsetty Seminole High School



meteorology, Machine Learning, weather patterns


Understanding upcoming weather patterns is extremely important for several different reasons in a community. For instance, weather patterns can hint at upcoming threats to a community such as hurricanes or tornados and give people a proper amount of time to prepare or evacuate the area. Also, weather patterns can give members of the agricultural field notice of rainfall the following day, week, or even month. This allows them to make better grounded decisions in order to produce better produce for their community. Today, numerical weather predictions models serve as the basis for weather predictions for several different weather news sources. However, the recent influx of Machine Learning has made few question whether machine learning models might be a better fit for weather predictions. Machine Learning consists of several different models and algorithms which all perform in different ways. The aim of this research is to compare three different machine learning models in both their accuracy and their progressive learning given more training data by using the field of meteorology as a testbed. The models were given several points of data and expected to return a value predicting whether it would rain the following day with varying amounts of training data. The three different models used were a Logistic Regression, Multilayer Perceptron, and Support Vector Machine models. After conducting the tests, it was found that the Multilayer Perceptron model predicted wielding the greatest accuracy. Furthermore, it was similarly found that the Multilayer Perceptron model had the greatest positive progression as more training data was inputted.


Numerical weather prediction (weather models) - National Weather Service. (n.d.).

How has weather forecasting changed over the past two hundred years?. American Geosciences Institute. (2016, November 17).

DeLoss, J. (2021, October 28). From research to real world: CSU atmospheric scientists develop heavy rainfall forecast tool used nationwide. Walter Scott, Jr. College of Engineering.

Manning, A. (2023, March 1). CSU machine learning model helps forecasters improve confidence in storm prediction. Walter Scott, Jr. College of Engineering.

Paniagua-Tineo, A. & Salcedo-Sanz, Sancho & Casanova, Carlos & Ortiz-García, Emilio & Cony, Marco & Hernández-Martín, E.. (2011). Prediction of daily maximum temperature using a support vector regression algorithm. Renewable Energy. 36. 3054-3060. 10.1016/j.renene.2011.03.030.

Ibrahim Gad & Doreswamy Hosahalli (2022) A comparative study of prediction and classification models on NCDC weather data, International Journal of Computers and Applications, 44:5, 414-425, DOI: 10.1080/1206212X.2020.1766769

Anjali, T., Chandini, K., Kadan, A., & Lajish, V.L. (2019). Temperature Prediction using Machine Learning Approaches. 2019 2nd International Conference on Intelligent Computing, Instrumentation and Control Technologies (ICICICT), 1, 1264-1268.

Cho, D., Yoo, C., Im, J., & Cha, D.-H. (2020). Comparative assessment of various machine learning-based bias correction methods for numerical weather prediction model forecasts of extreme air temperatures in urban areas. Earth and Space Science, 7, e2019EA000740.

natesilver538. (2014, December 4). Which city has the most unpredictable weather?. FiveThirtyEight.

Carylsue. (2017, May 31). What cities have the most unpredictable weather in the U.S.?. National Geographic Education Blog.

East Boston Weather History. Weather Underground. (n.d.).

McKinney, W. (2008). Pandas. pandas.

Heikkilä, M. (2023, July 27). Weather forecasting is having an AI moment. MIT Technology Review.,methods%2C%20and%20much%20more%20quickly.

Stillman, D. (2023, July 5). Should we trust ai to predict natural disasters?. The Washington Post.