Comprehensive comparison between artificial intelligence and multiple regression: prediction of Palmerston North’s temperature
Abstract New Zealand is a country of islands consisting of two prominent landmasses: the North Island and the South Island. It is the sixth-largest island country in the eastern part of Australia. The South Island is generally cooler than the North Island, resulting in a temperature range in the Sou...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2025-07-01
|
| Series: | Discover Sustainability |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s43621-024-00745-x |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract New Zealand is a country of islands consisting of two prominent landmasses: the North Island and the South Island. It is the sixth-largest island country in the eastern part of Australia. The South Island is generally cooler than the North Island, resulting in a temperature range in the South that lags behind that of the northern region. Palmerston North is located in the lower part of the North Island, along with the capital city, Wellington. In this study, we predict the temperature data for Palmerston North using initial temperature data from four neighboring cities: Napier, New Plymouth, Taupo, and Wellington. We then reduce the analysis to three cities by excluding Taupo. The temperature data spans from 1971 to 2019 for our investigation. Three alternative algorithms have been developed and applied to the dataset: an artificial neural network (ANN) with a single hidden layer, multiple regression, and supervised machine learning. We found that all three algorithms performed well, successfully predicting the desired temperature data. To evaluate the performance of these models, we analysed the errors between the true values and the predicted values for the study dataset, which included the mean squared error (0.00016, 0.00016, & 0.00012, respectively) and the modeling efficiency (0.9883, 0.9988, & 0.9991, respectively), among others. We found that multiple linear regression and supervised machine learning slightly outperformed the ANN. All algorithms were executed in Python 3.9.16 under the Anaconda platform. |
|---|---|
| ISSN: | 2662-9984 |