Skripsi

MODEL PREDIKSI MULTIPLE ANFIS-CART PERTUMBUHAN TANAMAN SORGHUM BERDASARKAN PEMBERIAN PUPUK ORGANIK PADA LAHAN PASANG SURUT.

---

Penilaian

---

The investigation into sorghum cultivation indicates that sorghum demands reduced fertilizer input, exhibits greater resilience to drought and unfavorable conditions, and possesses significant nutritional value. Sorghum serves as a food source and finds applications as animal feed and a raw material for bread production. Furthermore, it can use its biomass as an alternative energy source—this positions sorghum as a viable candidate for cultivation in Indonesia to bolster the national food diversification program. Organic fertilizers, which are local resources, can enhance the nutrient content in sorghum cultivation. The research utilizes the Adaptive Neural Fuzzy Inference System (ANFIS) model for forecasting the height, biomass weight, and seed weight of each sorghum plant head. Given the use of three types of organic fertilizers (chicken manure, cow dung, and vermicompost) and different combinations of fertilizer and dolomite lime, the study employs a Multiple ANFIS model, comprised of nine individual ANFIS models. Decision-making regarding sorghum plant growth is based on the predicted output parameters, and a Decision Tree model, precisely CART (Classification and Regression Trees), is employed. Hence, the innovation of this research is rooted in utilizing the Multiple ANFIS-CART approach for the predictive model. The dissertation outcomes reveal that the Multiple ANFIS model employed in forecasting sorghum plant growth comprises nine ANFIS models characterized by different types and degrees of membership functions. These variations are determined based on the accuracy level achieved for each organic fertilizer treatment and the predicted output parameters. The most accurate prediction in the Multiple ANFIS model is observed in forecasting seed weight per sorghum plant head, specifically when using chicken manure, resulting in a MAPE of 5.77%, MAE of 0.2994, and RMSE of 0.395. The K-Means Clustering algorithm's clustering process aims to provide labels for three predicted parameters, and the optimal result yields three clusters. The decision-making process using the CART decision tree is employed to determine the growth of sorghum plants classified as Very Good, Good, and Less Good. The accuracy of the CART model results is measured using a confusion matrix, indicating an accuracy of 96% when using the Gini index criteria, while the CART model using the entropy criteria achieves an accuracy level of 99%. Keywords: Prediction, Sorghum, Multiple ANFIS-CART, K-Means Clustering, Accuracy.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2407001251	T140127	T1401272024	Central Library (Referens)	Available

Detail Information

Series Title

-

Call Number

T1401272024

Publisher

Indralaya : Prodi Doktor Ilmu Teknik, Fakultas Teknik Universitas Sriwijaya., 2024

Collation

xvi, 96 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

621.390 7

Content Type

Text

Media Type

unmediated

Carrier Type

-

Edition

-

Subject(s)

Ilmu Teknik, Teknik Komputer

Specific Detail Info

-

Statement of Responsibility

MURZ

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment