Table 1 Summary of recent research studies of text-mining applications for financial predictions
From: Comprehensive review of text-mining applications in finance
Study | Datasets | Techniques/algorithms used | Evaluation parameters | Performance (on the basis of evaluation parameters) | References |
|---|---|---|---|---|---|
Stock index forecasting | German ad hoc announcements and stock indices (DAX, CDAX, and STOXX) | Lasso, ridge regression, elastic net, gradient boosting, random forest | Root mean squared error (RMSE) | DAX: 409.662 CDAX: 35.273 STOXX: 12.170 | Feuerriegel and Gordon (2018) |
Stock prices prediction | Apple’s stock prices, July 2017 | SVR | Comparison of kernels | RBF kernel performed best | Shah et al. (2018b) |
AZFinText system for stock market prediction | Financial news articles | SVR | Closeness, directional accuracy, simulated trading engine | Closeness: 0.04261 Directional accuracy: 57.1% Simulated trading: 2.06% return | Schumaker and Chen (2009) |
Financial time series forecasting | Security companies’ quarterly and annual reports | ARIMA, SVR | Mean absolute error (MAE), mean absolute per cent error (MAPE), RMSE | MAE: 1.07 MAPE: 23.06 RMSE: 1.34 | Wang et al. (2012) |
Stock price prediction | Online Flemish newspaper articles | SVM | Accuracy, AUC, return rate, Sharpe ratio | Varied results across the defined metrics for different techniques; suggestion for a new evaluation parameter | Junqué de Fortuny et al. (2014) |
Stock price prediction by automated news reading | Corporate announcements, stock price data | SVM, SVR | Accuracy, Squared correlation coefficient (R2) | Accuracy: 76.3% R2: 20.2% | Hagenau et al. (2013) |
Forex intraday trend prediction | News headlines, currency pair rates | J48 classifier, synchronous target feature reduction | Accuracy | 80% | Vijayan and Potey (2016) |
Mining critical indicators of stock market movements | Financial news articles | RF | Accuracy | 98.34% | Elagamy et al. (2018) |
Using news sentiments for stock price prediction | News articles related to the Nifty Pharma Index | Dictionary-based sentiment analysis model | Directional accuracy | 70.59% | |
Exchange rates prediction | News data, social media information by Market Psych | Multivariate linear regression, multilayer perceptron | Directional accuracy | 60.26% | Yusuuf and Shihabeldeen (2019) |