We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Modified Generalized Lasso for Variable Selection in Lag Distributed Modeling of Fresh Fruit Bunch Production from Oil Palm Plantations in Riau-Indonesia.
- Authors
Kurnia, Anang; Rahardiantoro, Septian; Oktarina, Sachnaz Desta; Anisa, Rahma; Nur Rahman, Nafisa Azzahra; Handayani, Dian
- Abstract
This research identified factors affecting the productivity of oil palm fresh fruit bunches (FFB) in metric tons per ha (hectare). Current research rarely includes spatial and temporal aspects, so we proposed the modified Generalized Lasso, which can be used in the lagdistributed regression by considering the adjacency of time lags and locations in the data. The modification is located in how the definition of the regression model and the penalty matrix in the Generalized Lasso, which considers the adjacencies between blocks and time lags. The method is applied on plantations in the Riau context in Indonesia. The oil palm management data used consists of 42 months of observations in 16 planting blocks spanning from 2020 to 2023. The response variable was the productivity of oil palm FFB, with predictor variables consisting of the number of rainy days, rainfall, the dosage of NPK fertilizer, and palm age. We compared our proposed method with standard Lasso. As a result, our proposed model obtained a smaller error value than the standard lasso models. It is indicated that the lag of the productivity variable and the lag of the number of rainy days influence the FFB productivity for almost all blocks.
- Subjects
RIAU (Indonesia : Province); INDONESIA; PALM oil industry; PALMS; INDEPENDENT variables; FRUIT; PLANTATIONS; OIL palm
- Publication
International Journal of Advances in Soft Computing & Its Applications, 2024, Vol 16, Issue 1, p1
- ISSN
2710-1274
- Publication type
Article
- DOI
10.15849/IJASCA.240330.01