Paper Type
Complete
Abstract
Word-of-mouth (WOM) communication may reflect investor sentiment and influence stock price dynamics. This study evaluates whether publicly available WOM predicts subsequent price movements and trend changes on the Warsaw Stock Exchange (WSE). Textual data were collected from Bankier.pl, including financial news and stock forum discussions, while historical price data were obtained from Investing.com. A rule-based framework was applied to assess the consistency between sentiment signals and subsequent price changes, as well as their relevance to identified trend reversals. Trends were detected using selected technical indicators, including the Average Directional Index (ADX), Simple and Exponential Moving Averages (SMA/EMA 10/21), and the Ichimoku Cloud. Sentiment analysis was conducted using a Polish-language sentiment classifier. The results indicate that WOM demonstrates predictive potential, with market reactions often occurring with a delay. However, the strength of the relationship depends on indicator parameterization and sentiment-model performance. Further research should expand data sources and develop a domain-specific Polish financial sentiment model.
Paper Number
1247
Recommended Citation
Szuta, Adam and Weichbroth, Pawel, "The Effect of Word-of-Mouth on Stock Price Movements: A Rule-based Approach" (2026). AMCIS 2026 Proceedings. 4.
https://aisel.aisnet.org/amcis2026/ent_system/sig_entsys/4
The Effect of Word-of-Mouth on Stock Price Movements: A Rule-based Approach
Word-of-mouth (WOM) communication may reflect investor sentiment and influence stock price dynamics. This study evaluates whether publicly available WOM predicts subsequent price movements and trend changes on the Warsaw Stock Exchange (WSE). Textual data were collected from Bankier.pl, including financial news and stock forum discussions, while historical price data were obtained from Investing.com. A rule-based framework was applied to assess the consistency between sentiment signals and subsequent price changes, as well as their relevance to identified trend reversals. Trends were detected using selected technical indicators, including the Average Directional Index (ADX), Simple and Exponential Moving Averages (SMA/EMA 10/21), and the Ichimoku Cloud. Sentiment analysis was conducted using a Polish-language sentiment classifier. The results indicate that WOM demonstrates predictive potential, with market reactions often occurring with a delay. However, the strength of the relationship depends on indicator parameterization and sentiment-model performance. Further research should expand data sources and develop a domain-specific Polish financial sentiment model.
Comments
SIG ENTSYS