Pemodelan dalam Prediksi Pembentukan Harga Terendah Baru pada Emiten Pasar Saham Indonesia

Firdaus, Nelwan Topan (2026) Pemodelan dalam Prediksi Pembentukan Harga Terendah Baru pada Emiten Pasar Saham Indonesia. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6032241010-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy

Abstract

Volatilitas pasar saham Indonesia meningkat seiring bertambahnya investor ritel dan tingginya sensitivitas terhadap perubahan makroekonomi serta ketidakpastian global. Penurunan harga pada banyak emiten mencerminkan tekanan pasar yang dapat memicu krisis finansial. Pendekatan kuantitatif dinilai lebih andal dalam menganalisis penurunan harga dibandingkan informasi spekulatif di media. Berdasarkan kondisi tersebut, penelitian ini bertujuan membangun kerangka analitis untuk memodelkan keterkaitan variabel ekonomi dengan pasar saham serta menghasilkan estimasi prediktif atas dinamika ekonomi. Penelitian ini menggunakan pendekatan kuantitatif berbasis data aktual dengan dua model deret waktu, yakni ARIMAX untuk menangkap pola historis harga saham dan pengaruh variabel eksogen, serta MIDAS-OLS dan MIDAS-Logit untuk mengintegrasikan variabel berfrekuensi campuran harian-bulanan dan mendeteksi probabilitas kejadian breakdown ekstrem tanpa agregasi data. Pemilihan ordo ARIMAX dilakukan melalui Grid Search berbasis nilai Akaike Information Criterion (AIC) terendah, sementara performa model dievaluasi menggunakan MAPE, MAE, MSE, dan RMSE pada ARIMAX, serta ROC-AUC, Confusion Matrix, dan Lead Time pada MIDAS-Logit. Hasil penelitian menunjukkan bahwa ARIMAX menghasilkan tingkat akurasi prediksi terbaik pada sektor finance dengan nilai MAPE terendah dan stabil. Model MIDAS-OLS mampu mengidentifikasi variabel makro yang signifikan dalam memengaruhi pergerakan harga saham, dengan sensitivitas paling kuat pada sektor consumer cyclical. Sedangkan MIDAS-Logit efektif memberikan sinyal awal probabilitas breakdown melalui nilai ROC-AUC yang dapat diandalkan sebagai dasar sistem peringatan dini, dan yang memiliki performa tertinggi pada sektor consumer non-cyclical dengan nilai ROC-AUC sebesar 0,8003 dan 0,7045. Dengan demikian, kombinasi ARIMAX dan MIDAS merupakan pendekatan yang relevan untuk mendukung pengambilan keputusan strategis regulator dan pelaku pasar dalam mitigasi risiko krisis.
==================================================================================================================================
The volatility of the Indonesian stock market has increased alongside the growing number of retail investors and heightened sensitivity to macroeconomic changes and global uncertainty. Price declines across numerous listed firms reflect market pressure that may trigger financial crises. Quantitative approaches are considered more reliable for analyzing price downturns than speculative information disseminated through the media. In this context, this study aims to develop an analytical framework to model the relationship between economic variables and the stock market, as well as to generate predictive estimates of economic dynamics. This research employs a quantitative approach using actual market data and two time-series modeling techniques, namely ARIMAX to capture historical stock price patterns and the influence of exogenous variables, and MIDAS-OLS and MIDAS-Logit to integrate mixed-frequency daily–monthly variables and detect the probability of extreme breakdown events without data aggregation. ARIMAX order selection is conducted through a Grid Search based on the lowest Akaike Information Criterion (AIC) value, while model performance is evaluated using MAPE, MAE, MSE, and RMSE for ARIMAX, and ROC-AUC, Confusion Matrix, and Lead Time for MIDAS-Logit. The findings indicate that the ARIMAX model achieves the highest predictive accuracy in the finance sector, as reflected by the lowest and most stable MAPE values. The MIDAS-OLS model effectively identifies significant macroeconomic variables influencing stock price movements, with the strongest sensitivity observed in the consumer cyclical sector. Meanwhile, the MIDAS-Logit model provides reliable early signals of breakdown probability through robust ROC-AUC performance, achieving its highest values in the consumer non-cyclical sector (ROC-AUC of 0.8003 and 0.7045). Accordingly, the integration of ARIMAX and MIDAS constitutes a relevant methodological framework to support strategic decision-making by regulators and market participants in mitigating financial crisis risk.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	ARIMAX, makro ekonomi, MIDAS, peringatan dini, volatilitas pasar saham, early warning system, macroeconomics, stock market volatility.
Subjects:	H Social Sciences > HB Economic Theory > Economic forecasting--Mathematical models.
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User:	Nelwan Topan Firdaus
Date Deposited:	27 Jan 2026 03:14
Last Modified:	27 Jan 2026 03:14
URI:	http://repository.its.ac.id/id/eprint/130152

Actions (login required)

View Item