A Machine-Learning Approach to Time Series Forecasting of Temperature

You are here: Homepage > Search

A Machine-Learning Approach to Time Series Forecasting of Temperature

Pant, Janmejay / Sharma, Rakesh Kumar / Juyal, Amit / Singh, Devendra / Pant, Himanshu / Pant, Puspha

In the modern world, weather forecasting is an essential application. The forecasts can help us reduce weather- related losses. The need for a massive data and highly computationally intensive parameterization procedure can be eliminated or reduced by the use of machine learning and deep learning algorithms for forecasting. This research work intends to forecast the temperature of three cities (Dehradun, Mukteshwar and Pantnagar) of Uttarakhand using emerging time series model AR/MA (Auto Regressive Integrated Moving Average). The results of this study prove that using the auto ARIMA produces very less MAPE (Mean Absolute Percentage Error) score for all testing data of all three regions. Our used model produces 8.45%,9.65% and 5.64% mean absolute percentage error in temperature data for Dehradun, Mukteshwar and Pantnagar respectively. Hence this paper explains briefly how different parameters can be used to formulate the AR/MA model to predict temperature. MAPE (Mean Absolute Percentage Error) indicates that auto AR/MA model yields excellent results.

Access

Check access

Check availability in my library

Order at Subito €

Export, share and cite

Document information

Title :

A Machine-Learning Approach to Time Series Forecasting of Temperature

Contributors:

Pant, Janmejay (author) / Sharma, Rakesh Kumar (author) / Juyal, Amit (author) / Singh, Devendra (author) / Pant, Himanshu (author) / Pant, Puspha (author)

Published in:

2022 6th International Conference on Electronics, Communication and Aerospace Technology ; 1125-1129

Publication date :

2022-12-01

Size :

1445983 byte

ISBN :

978-1-6654-8271-4 , 978-1-6654-8270-7

DOI :

https://doi.org/10.1109/ICECA55336.2022.10009165

Type of media :

Conference paper