FUNDAMENTAL ANALYSIS AND SHORT-TERM STOCK PRICE FORECASTING USING MULTIPLE REGRESSION ANALYSIS AND ADAPTIVE NEURO FUZZY INFERENCE SYSTEM (CASE STUDY: PT TIMAH TBK)

• Main Author: Saputra, Aleo
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/85345
• Institution: Institut Teknologi Bandung

PT Timah Tbk is a tin mining company that conducted its Initial Public Offering (IPO) on October 19, 1995. Selling shares to the public is a strategic way to raise additional funds, which can be used 1o streamline the company's operations or accelerate expansion activities. During an IPO, a company obtains new capital by raising funds from the public through the sale of shares in the primary market, while shareholders can benefit from trading shares through capital gains in the secondary market and dividends distributed by the company regularly. Shares of a company with strong performance are ofien in high demand among investors, leading to an increase in stock price. Since mid-2022, the share price of TINS.JK has shown a downward trend through early 2024. This decline is considered a sign of decreased company performance, also reflected by a drop in net profits since the first quarter of 2022. If this condition persists, it could lead to uncertainty in capital gains and dividends, potentially causing losses for investors. To anticipate larger losses in the future, investors need to evaluate their position regarding TINS.JK shares. This research employs fundamental analysis to examine the relationship between the company's financial ratios and stock price formation. Fundamental analysis is conducted using the Multiple Regression Analysis (MRA) method. Meanwhile, 1o predict the movement direction of TINS.JK Shares, forecasts using the Adaptive Neuro Fuzzy Inference System (ANFIS) method and Multiple Regression Analysis (MRA) are compared. The analysis with MRA shows that only the variables earning per share (EPS) and price to book value (PBV) significantly affect the average share price (ASP). In contrast, forecasting with ANFIS yields a lower Mean Absolute Percentage Error (MAPE) 0f 7.61% compared t0 8.64% with MRA.