Temperature Forecasting and Analysis Using Linear and Timeseries Models

• Author(s): Rishabh Singh ; Shreyash Sherigar ; Dr. Santosh Singh
• Paper ID: 1705460
• Page: 78-84
• Published Date: 03-02-2024
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 7 Issue 8 February-2024

Abstract

Effective weather forecasting is a cornerstone in various sectors, ranging from agriculture and disaster management to climate research. Accurate predictions facilitate proactive decision-making and the mitigation of potential risks. This research paper presents a comprehensive investigation aimed at enhancing weather forecasting through a systematic exploration of linear and time series models. The study commences with meticulous data collection from diverse sources, encompassing weather stations, satellite observations, oceanic sensors, and historical records. Subsequent data preprocessing ensures data quality and integrity. Various climate parameters, including solar radiation, wind direction, and carbon emissions, are examined to refine forecasting accuracy and understand their intricate interactions. Christopher Vu et al [4]. and Mehmet et al. [3] in their research have proven the effectiveness of Linear and Time-series models building upon that, the initial modeling phase employs linear models, featuring linear regression and L1 regularization, known for their simplicity and effectiveness as our focused region is Mumbai metropolitan which does not experience extreme temperature changes. Moving forward, the research delves into time series models, notably ARIMA and its seasonal variation, SARIMA. These models are introduced to capture the temporal dynamics and dependencies inherent in climate data. The research highlights the suitability of specific models for different climate parameters. Linear models prove valuable for stable, less-variable parameters, while time series models excel in capturing intricate temporal dependencies. The shift from linear to time series models results in significantly enhanced accuracy in temperature predictions, emphasizing the need for tailored methodologies in data-driven projects. This paper contributes to the realm of weather forecasting and analysis by offering a structured approach to model selection, dataset preparation, and the interpretation of complex climate dynamics. It underscores the critical role of precise forecasting in promoting environmental sustainability and well-informed decision-making.

Keywords

ARIMA, Linear Regression, Mumbai Metropolitan Region, Time Series, Weather Forecasting

Citations

IRE Journals:
Rishabh Singh , Shreyash Sherigar , Dr. Santosh Singh "Temperature Forecasting and Analysis Using Linear and Timeseries Models" Iconic Research And Engineering Journals Volume 7 Issue 8 2024 Page 78-84

IEEE:
Rishabh Singh , Shreyash Sherigar , Dr. Santosh Singh "Temperature Forecasting and Analysis Using Linear and Timeseries Models" Iconic Research And Engineering Journals, 7(8)