Forecasting Red Chilli Plant Growth using Time Series Method With Long Short-Term Memory Model

Abstract

The growth of red chilli plants is a horticultural commodity whose growth is highly determined by environmental elements, as a result, it is very crucial to make predictions to help more effective agricultural planning. This study aims to examine the ability of the Long Short-Term Memory (LSTM) model in predicting the growth of red chilli plants (Capsicum annuum L.) according to 4 main parameters, namely stems, branches, leaves, and grains. The data used are red chilli plant growth data obtained from plantations located in Deli Serdang Regency, precisely in Namorambe District, namely Jatikusuma Village, over a period of 63 days and analyzed using the time collection method. The example provides high prediction accuracy for stem parameters (R² = 0.9796), branches (R² = 0.9618), and leaves (R² = 0.9489), but slightly low in fruit (R² = 0.8807) due to hyperbolic fluctuations. The consequences show the potential of LSTM in helping red chilli cultivation through better planning, green aid control, and early detection of growth anomalies. This study also demonstrates an integrative approach to four plant growth parameters using a single LSTM instance.