Copyright@ REST Publisher 9 Senthilkumar Meyyappan.et.al / Journal on Electronic and Automation Engineering, 2(4) December 2023, 9-18 Journal on Electronic and Automation Engineering Vol: 2(4), December 2023 REST Publisher; ISSN: ISSN: 2583-6951 (Online) Website: https://restpublisher.com/journals/jeae/ DOI: https://doi.org/10.46632/jeae/2/4/2 Improving Weather Forecasting Accuracy Using Machine Learning *Senthilkumar Meyyappan, A. Bharath Naik, A. Uma Sai, Ch. Keerthi Nalla Malla Reddy Engineering College, Hyderabad, India. *Corresponding author Email: kathir_senthil@yahoo.co.in Abstract: Weather forecasting has several applications in our daily lives, ranging from agriculture to event planning. Previous weather forecasting models relied on a complex combination of mathematical instruments, which was insufficient to achieve a higher categorization rate. We offer fresh revolutionary approaches for estimating monthly rainfall using machine learning algorithms in this study. Weather forecasts are created by gathering quantitative information about the current state of the atmosphere. Machine learning algorithms may learn complicated mappings from inputs to outputs using only samples and with little effort. The dynamic nature of the atmosphere makes accurate weather prediction challenging. The fluctuation in weather conditions in previous years must be used to anticipate future weather conditions. It is extremely likely that it will match within the next two weeks of the preceding year. We proposed using linear regressions with the Random forest algorithm to forecast weather using characteristics such as temperature, humidity and wind. It will forecast weather based on prior records thus, this prediction will be accurate. 1. INTRODUCTION Weather forecasting is primarily concerned with predicting weather conditions in the future. Weather predictions provide essential information about the weather in the future. Weather forecasting methodologies range from relatively simple observation of the sky to highly complicated computerized mathematical models. Weather forecasting is critical for a variety of purposes. Climate monitoring, drought detection, severe weather prediction, agricultural and production, energy industry planning, aviation industry planning, communication, pollution dispersal, and so on are some of them. Weather Underground's free tier API web service will be utilized to acquire the data for this series. Since 2015, I've been interacting with the API via the requests library to retrieve weather data for the city of Lincoln, Nebraska. Once acquired, the data must be processed and aggregated into a format suitable for data analysis before being cleaned. The planet is a very complicated area that is suffering from climate change. It is critical to predict precise weather without error, to ensure security and mobility, as well as a safe daily operation. Weather forecasting is built by collecting massive amounts of data, which makes machine learning an indispensable tool, by employing some back testing methods and algorithms to generate an accurate prediction of the weather. We will evaluate the methods through a series of experiments that will emphasize their performance and worth. The addition of ML to weather forecasting provides a significant advantage for the prediction make it more accurate. Weather forecasting is a very effective application of science for the benefit of society. It can be used for Aerospatiale, agriculture, sport, musical events, maritime, renewable energy, aviation, and forestry, where it is critical to know what the weather will be. Finding the highest and minimum temperature is one of the linear regression methods that yields an accurate answer.  Collecting the data.  Exploring and preparing data  Data preparation- creating random training and test data.  Training model on data.  Evaluating model performance.  Improving model performance.