Pemodelan Indeks Ketahanan Pangan Beras di Jawa Timur Menggunakan Regresi Non Linear dan Algoritma Genetika

Hastuti, Desi Puji (2017) Pemodelan Indeks Ketahanan Pangan Beras di Jawa Timur Menggunakan Regresi Non Linear dan Algoritma Genetika. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
1315105021-Undergraduate_Theses.pdf - Published Version

Download (2MB) | Preview

Abstract

Negara dengan sumber ekonomi cukup memadai dapat mengalami kehancuran karena tidak mampu memenuhi kebutuhan pangan bagi penduduknya. Ukuran ketersediaan pangan suatu daerah dapat digambarkan oleh Indeks Ketahanan Pangan. Adanya hubungan yang tidak linear antara indeks ketahanan pangan dengan beberapa variabel yang diduga mempengaruhinya menyebabkan regresi model linear kurang tepat untuk digunakan. Oleh karena itu dalam memodelkan indeks ketahanan pangan beras di Jawa Timur digunakan metode regresi linear dan regresi non linear dimana optimasi estimasi parameter dilakukan menggunakan Algoritma Levenberg-Marquardt dan Algoritma Genetika. Dari ketiga model tersebut akan dibandingkan berdasarkan kriteria kebaikan model yaitu nilai RMSE. Hasil analisis menunjukkan bahwa rasio ketersediaan beras minimum pada Kota Surabaya sebesar 0,058 yang disebabkan karena jumlah konsumsi lebih banyak daripada jumlah produksi beras yang tersedia. Pemodelan menggunakan regresi linear telah memenuhi semua asumsi, tetapi nilai R2 rendah. Dari hasil perbandingan model berdasarkan kriteria RMSE didapatkan hasil bahwa model yang menggunakan algoritma genetika sebagai metode estimasi parameter model regresi non linear merupakan model yang terbaik untuk memodelkan indeks ketahanan pangan beras di Jawa Timur karena memiliki nilai RMSE terkecil. =================================================================================== Countries with adequate economic resources can experience the devastation due to not being able to meet the needs of food for its population. The size of a region's food availability can be described by the Food Security Index. The existence of a non linear relationship between food security index with multiple variables that allegedly affected causing linear regression model less appropriate for use. Therefore in the model index of rice food security in East Java to use linear regression method and the non linear regression where optimization parameter estimation is done using Levenberg-Marquardt Algorithm and Genetic Algorithm. Of these three models will be compared based on the criteria of value model goodness namely RMSE. The results of the analysis showed that the ratio of the minimum availability of rice in Surabaya city of 0.058 caused due to the amount of consumption is more than the amount of rice production are available. Using linear regression modeling has fulfilled all the assumptions, but the value of R2 is low. From the results of the comparison of the model based on the criteria of RMSE obtained the results that the model that uses genetic algorithms as a method of parameter estimation of a non linear regression model is the best model to model index of rice food security in East Java because it has the smallest RMSE values.

Item Type: Thesis (Undergraduate)
Additional Information: RSSt 519.36 Has p
Uncontrolled Keywords: Algoritma Genetika, Indeks Ketahanan Pangan Beras, Regresi Non Linear
Subjects: H Social Sciences > HA Statistics
Q Science > QA Mathematics > QA278.2 Regression Analysis
Divisions: Faculty of Mathematics and Science > Statistics > (S1) Undergraduate Theses
Depositing User: Desi Puji Hastuti
Date Deposited: 25 Oct 2017 07:19
Last Modified: 05 Mar 2019 08:38
URI: http://repository.its.ac.id/id/eprint/47784

Actions (login required)

View Item View Item