ONLINE NEWS SENTIMENT, TRADING VOLUME, INFLATION, AND INVESTOR SENTIMENT IN INDONESIA

Abstract

Purpose: This study examines the dynamic relationships among online news sentiment, trading volume, inflation, and investor sentiment in the Indonesian market.

Research Methodology: The study combines Naïve Bayes classification of online financial news with time-series analysis using a vector autoregression (VAR), Granger causality tests, impulse response analysis, and short-horizon forecasting. The reported observation period is 2021–2024.

Results: In the market-sentiment equation, lagged inflation and lagged market sentiment are statistically significant; lagged online news sentiment and trading volume are not. The reported Granger tests indicate that inflation predicts market sentiment, market sentiment and trading volume predict one another, and trading volume predicts online news sentiment.

Limitations: The analysis covers 2021–2024 and reports an adjusted R-squared of 0.124759 for the market-sentiment equation. The limited explanatory power and the period-specific context constrain generalization.

Contribution: This study contributes empirical evidence on the dynamic interactions among machine-classified online news sentiment, investor sentiment, trading volume, and inflation in the Indonesian capital market. By combining Naïve Bayes sentiment classification with a VAR framework, it examines how information-based and market-based indicators relate over time in an emerging-market context.