Work

Tropical climate prediction method combining random forest and feature fusion

Year: 2025

Type: article

Abstract: Tropical cyclones pose significant threats to coastal populations, causing destruction and loss of life. Precisely forecasting the frequency and arrival dates is still a challenge. This research presents a technique for feature extraction and integration using a random forest (RF) model with a cascaded convolutional neural network. The approach combines different meteorological maps and uses a feature fusion technique to improve prediction accuracy. The RF model is optimized by a grid search algorithm. The results show that the proposed model outperforms conventional models to achieve a mean absolute error of 0.48 and a mean relative error of 14.14%.

Source: International Journal of Low-Carbon Technologies

Author: Guotao Liu

Institution: University College London

Language: English

Topic: Tropical and Extratropical Cyclones Research

Subfield: Atmospheric Science

Field: Earth and Planetary Sciences

Domain: Physical Sciences

SDGs: Climate action, Industry, innovation and infrastructure

Open Access status: gold

Estimated APC paid: $2,886

OpenAlex ID: W4406864855