<![CDATA[This study developed a web-based decision support assistant for sectoral stock technical analysis using Walk-Forward Analysis and Large Language Model-based explanations. The system was designed to identify the best-performing technical indicator for each sector based on historical out-of-sample evaluation and to present the results in an interpretable form. The study evaluated Moving Average Crossover, Moving Average Convergence Divergence, and Relative Strength Index using daily historical data from 40 stocks representing the Financial, Technology, Industrial, and Energy sectors on the Indonesia Stock Exchange. The evaluation applied a fixed-length rolling scheme consisting of a six-month in-sample period, a three-month out-of-sample period, and a three-month window shift. Signal performance was assessed using Average Forward Return at T+1, T+3, T+5, and T+10 trading days, with Directional Accuracy as the primary metric. The results showed that Relative Strength Index achieved the highest observed out-of-sample performance in the Financial sector with a Directional Accuracy of 51.52%, while Moving Average Crossover achieved the highest performance in the Technology, Industrial, and Energy sectors with 50.00%, 57.14%, and 71.43%, respectively. The system presented sector-based indicator results, BUY, SELL, or HOLD signals, performance metrics, price visualizations, explanatory narratives, and PDF reports. All 20 predefined Black Box Testing scenarios were completed successfully, User Acceptance Testing achieved a success rate of 99.83%, and the System Usability Scale produced a score of 78.63, which was categorized as Good, Grade B+, and Acceptable. The findings indicate that technical indicator performance varies across sectors and evaluation periods, supporting the use of chronological out-of-sample validation in sector-specific technical analysis.]]>
Copyrights © 2026
