Cahyaningtyas, Ratih (2019) Peramalan Harga Saham Menggunakan Metode Artificiial Neural Network (Studi Kasus: PT Aneka Tambang Tbk). Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
05211540000101-Undergraduate_Thesis.pdf Download (2MB) | Preview |
Abstract
Aktivitas penanaman modal menjadi hal yang menantang bagi investor. Faktor-faktor mempengaruhi harga saham akan ada ketidakpastian yang tidak bisa dihindari di pasar saham. Peramalan harga saham diharapkan dapat membantu dalam pengambilan keputusan finansial. ANN merupakan metode yang meniru sistem kerja jaringan syaraf. Metode ANN dapat mempelajari pola atau hubungan dari data itu sendiri dan dianggap sebagai pendekatan yang berguna untuk mengatasi masalah peramalan deret waktu pada harga saham yang tidak linier. Variabel yang digunakan dalam penelitian ini adalah harga saham penutupan harian dan data kurs tukar IDR-USD. Berdasarkan hasil Tugas Akhir, model ANN metode Backpropagation (BPNN) menghasilkan peramalan harga saham PT Aneka Tambang Tbk dengan baik. Model BPNN terbaik periode t-7 yaitu model 14-23-1. Hasil analisa menunjukkan bahwa model BPNN periode t-7 mempunyai nilai MSE 328,4855661 dan nilai MAPE 1,684522%. Ketika model BPNN dibandingkan dengan model RNN, model RNN memberikan nilai akurasi yang lebih baik dari model BPNN. Model terbaik RNN adalah model 2-4-1 dengan nilai MSE 301,9852581 dan nilai MAPE 0,519259%. =================================================================================================================================
Investment activities are challenging for investors. Factors that influence stock prices will always be there and want to be avoided in the stock market. Stock prices forecasting is expected to help in making financial decisions. ANN is a method that mimics a neural network working system. The ANN method can support patterns or relationships from the data itself and is considered as useful to overcome the problem of time series forecasting at non-linear stock prices. The variables used in this study are the daily closing stock price and the IDR-USD exchange rate. Based on the results, the ANN method Backpropagation (BPNN) results is good in forecasting of the stock price of PT Aneka Tambang Tbk. The best BPNN model for the t-7 period is models 14-23-1. The results of the analysis show that the period t-7 BPNN model has MSE value 328.4855661 and MAPE value of 1.684522%. When the BPNN model is compared to the RNN model, the RNN model provides better accuracy values than the BPNN model. The best model of RNN is the 2-4-1 model with the value of MSE 301.9852581 and the MAPE value of 0,519259%.
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSI 006.32 Cah p-1 2019 |
| Uncontrolled Keywords: | Data Historis, Kurs IDR-USD, Recurrent Neural Network, Backpropagation |
| Subjects: | T Technology > T Technology (General) T Technology > T Technology (General) > T174 Technological forecasting |
| Divisions: | Faculty of Information Technology > Information System > 57201-(S1) Undergraduate Thesis |
| Depositing User: | Ratih Cahyaningtyas |
| Date Deposited: | 22 Mar 2024 03:03 |
| Last Modified: | 22 Mar 2024 03:03 |
| URI: | http://repository.its.ac.id/id/eprint/64423 |
Actions (login required)
 |
View Item |