Mahistha, Dvitiya Srestha Prajna (2018) Prediksi Sebaran Jenis Ikan dengan Menggunakan Jaringan Saraf Tiruan - Algoritma Genetika. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
02311440000104-Undergraduate_Theses.pdf - Accepted Version Download (2MB) | Preview |
Abstract
Badan Informasi Geospasial (BIG) menyatakan bahwa garis pantai Indonesia memiliki total sepanjang 99.093 kilometer, namun terdapat permasalahan dalam dunia kelautan dan perikanan. Permasalahannya adalah informasi keberadaan ikan masih belum ada, sehingga dibutuhkan suatu perancangan prediksi untuk menginformasikan adanya keberadaan ikan. Tujuan penelitian ini adalah memperoleh algoritma prediktor terbaik untuk penentuan keberadaan ikan dan menghitung prediksi berat 3 jenis ikan (ikan biji nangka, grobyak, dan ikan cucut) dengan menggunakan algoritma prediktor terbaik di pantai Kenjeran secara real time. Penelitian ini diusulkan ada metode untuk memprediksi keberadaan ikan, yaitu JST dan JST-AG. Data sekunder yang didapatkan berasal dari NOAA, BMKG, dan PPN Brondong. Data sekunder independen dan dependen diuji missing data, outlier, dan korelasi. Kemudian JST dan JST-AG dirancang arsitekturnya. Arsitektur yang terbaik (RMSE terkecil) digunakan untuk memprediksi real time di pantai Kenjeran. Arsitektur terbaik didapatkan pada JST-AG dengan input berupa suhu permukaan laut, salinitas, dan curah hujan. Gabungan JST-AG dengan hidden node yang dipasang sebanyak 3, 1, dan 9 unit mampu menurunkan RMSE masing-masing ikan biji nangka, grobyak, dan cucut sebesar 351,88 kg; 37,99 kg; dan 1118,53 kg. Hasil berat prediksi ikan pada tanggal 1 Juli 2018 di pantai Kenjeran adalah ikan biji nangka, grobyak, dan cucut masing-masing sebanyak 0 kg; 3986,05 kg; dan 0 kg.
========================================================================================================
The Geospatial Information Agency (BIG) stated that the coastline of Indonesia has 99,093 kilometers, but there is problem in the world of marine and fisheries. The problem is the presence of fish information is not there, so it takes a prediction design to inform the existence of fish. The aim of this research is to obtain best predictor algorithm for determining fish existence and calculate the prediction of 3 species of fish (biji nangka, grobyak, and cucut) by using the best predictor algorithm at Kenjeran beach in real time. This research proposed there are methods to predict the existence of fish, namely JST and JST-AG. Secondary data obtained from NOAA, BMKG, and PPN Brondong. Independent secondary data and dependent tested missing data, outlier, and correlation. Then JST and JST-AG designed the architecture. The best architecture (the smallest RMSE) is used to predict real time on the Kenjeran beach. The best architecture is found in JST-AG with input in the form of sea surface temperature, salinity, and rainfall. Combined JST-AG with hidden nodes of 3, 1, and 9 units are able to decrease RMSE of each biji nangka, grobyak, and cucut fish of 351,88 kg; 37.99 kg; and 1118.53 kg. The results of fish prediction weight on July 1, 2018 at Kenjeran beach are jackfruit, grobyak and cucut fish of 0 kg each; 3986.05 kg; and 0 kg.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSF 006.3 Mah p-1 3100018077094 |
| Uncontrolled Keywords: | algoritma genetika, hidden node, jaringan saraf tiruan, real time, RMSE, genetic algorithm, hidden node, neural network, real time |
| Subjects: | Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods. T Technology > T Technology (General) > T174 Technological forecasting |
| Divisions: | Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | Dvitiya Srestha Prajna Mahistha |
| Date Deposited: | 16 Oct 2020 00:24 |
| Last Modified: | 16 Oct 2020 00:24 |
| URI: | http://repository.its.ac.id/id/eprint/55683 |
Actions (login required)
 |
View Item |