System for Determining Plant Types Based on Weather Characteristics and Soil pH Using Artificial Intelligence

Authors

  • Trisakti Akbar Universitas Negeri Makassar
  • Satria Gunawan Zain Universitas Negeri Makassar
  • Andi Baso Kaswar Universitas Negeri Makassar
  • Jumadi Mabe Parenreng Universitas Negeri Makassar

DOI:

https://doi.org/10.31763/iota.v5i2.902

Keywords:

k-nearest neighbor, long short-term memory, k-nn model, accuracy, mean absolute error

Abstract

This research implements the Long Short-Term Memory (LSTM) algorithm for weather forecasting using minimum temperature, maximum temperature, average temperature, air humidity, rainfall, and solar radiation values over the past 30 days. The output consists of forecasts for average temperature, air humidity, rainfall, and solar radiation for the next 30 days. The LSTM model output and soil pH are used to determine plant types using the K-Nearest Neighbor (K-NN) algorithm. Based on the LSTM model testing results, the minimum temperature feature achieved a Mean Absolute Error (MAE) of 0.0078, a maximum temperature of 0.0054, an average temperature of 0.009, air humidity of 0.0099, rainfall of 0.0042, and solar radiation of 0.0208. For the K-NN model, an accuracy of 98% was obtained.

Downloads

Published

2025-05-09

Issue

Section

Artificial Intelligence