Table 3.
Summary of the selected literature that incorporated online news data to forecast.
| Category | Research object | Author (Year) | Data set | Model | Evaluation criteria |
|---|---|---|---|---|---|
| Industry market forecasting | Tourism demand forecasting | Park et al. (2021) | China Daily and CNN online news | STM- SARIMAX | MAE, MAPE, RMSE, RMSPE,RI |
| Önder et al. (2019) | Online news | R-MIDAS | RMSE, MAE, MAPE, SMAPE, Theil's U | ||
| Stock returns forecasting | Narayan (2019) | Oil price news | OCR-Time series regression | MSFE, R2 | |
| Agricultural futures prices forecasting | Li et al. (2022) | Agricultural futures online news headlines | DP-Sent-LDA- BiLSTM-SVR/RF/BPNN | MAPE, RMSE | |
| Crude oil price forecasting | Li, Shang, and Wang (2019) | Crude oil news headlines | LDA–CNN–RFE-SVR/RF | MAE, RMSE | |
| Oil futures returns and volatility forecasting | Li, Jiang, Li, and Wang (2021) | Crude oil news headlines | VMD-BiLSTM | MSE, RMSE, MAE, HMSE, HMAE | |
| Oil market forecasting | Wu, Wang, Wang, and Zeng (2021) | Online oil news | CNN-BPNN/MLR/SVM/LSTM/RNN | MAE, MAPE, RMSE | |
| Macroeconomic indicators forecasting | Economic indicators forecasting | Song and Shin (2019) | Online economic news articles | Lexicon approach-MA/ARMA | RMSE, MAE,R2 |
| Macroeconomic forecasting | Tilly, Ebner, and Livan (2021) | Newspaper articles | BiLSTM-AR | RMSE, DM | |
| Economic index forecasting | Shapiro, Sudhof, and Wilson (2022) | Economic and financial newspaper articles | NLP text sentiment analysis techniques-regression | Adjusted R2 | |
| Economic forecasting | Ardia, Bluteau, and Boudt (2019) | Economic newspaper articles | Lexicons-Penalized least squares regression- | RMSFE, MAFE | |
| Elections and Politics | Elections forecasting | Fronzetti Colladon (2020) | Voting events online news | Social network analysis/text mining/Semantic brand score | MAPE, MAE |
| Public health and disease surveillance | Health-care stock prices forecasting | Shynkevich, McGinnity, Coleman, and Belatreche (2016) | Stocks news articles | GICS-MKL | Accuracy, Return |
| Economic policy | Economic policy uncertainty forecasting | Tobback, Naudts, Daelemans, Junquéde Fortuny, and Martens (2018) | Economic news | SVM-OLS | RMSPE |
Notes: Models: Structural topic models (STM); Ordinary least squares regression (OLS); Restricted MIDAS (R-MIDAS); Optical character recognition (OCR); Dependency parsing sentence latent dirichlet allocation (DP-Sent-LDA); Bi-directional long short-term memory (BiLSTM); Latent dirichlet allocation (LDA); Convolutional neural network (CNN); Recursive feature elimination (RFE); Moving average (MA); Global industry classification standard (GICS); Multiple kernel learning (MKL). Performance measure: Symmetric mean absolute percentage error (SMAPE); Mean squared forecast errors (MSFE); Mean Square Error (MSE); Heteroscedasticity adjusted mean squared error (HMSE); Heteroscedasticity adjusted mean absolute error (HMAE); Root mean squared forecast error (RMSFE); Mean absolute forecast error (MAFE).