Table 2 Summary of recent research studies on some text-mining applications for banking
From: Comprehensive review of text-mining applications in finance
Study | Datasets | Techniques/algorithms used | Evaluation parameters | Performance (on the basis of evaluation parameters) | References |
|---|---|---|---|---|---|
Topic extraction for Italian banks | Data from Twitter API | Topic modelling, LDA | Graphical representations | – | Krstić et al. (2019) |
Bank risk report quality assessment | Annual risk reports of German banks | Regressions | MAE, RMSE | MAE: 0.0338 RMSE: 0.0436 | Fritz and Tows (2018) |
Anti-money laundering reduction | Money laundering cases from the Internal Revenue Service | SAS visual text analytics | Defining use cases for text analytics | – | Cook and Herron (2018) |
Analysis of central bank documents | Bank of Italy documents | Vector space model, sentiment analysis | Automated readability index (ARI), Formality score | ARI: 12–15, Formality score: 70–75 | Bruno (2016) |
Bank service quality assessment | Client reviews from www.banki.ru | Rule-based classifier | Precision, recall, F-measure | Average F-measure: 0.86 | Bidulya and Brunova (2016) |
Bank failure prediction | Bank 10-K reports | SVM | Accuracy, precision, recall, F1 score | Accuracy: 84.34% Precision: 10.81 Recall: 40 F1: 17.02 | Gupta et al. (2016) |
Risk detection in the banking system | CEO letters and outlook sections in banks’ annual reports | NB, SVM | Accuracy, precision, recall | Accuracy (best out of all): 79.2% | Nopp and Hanbury (2015) |
Distress prediction by evaluating sentiments | Annual reports and financial statements | Feed forward neural network (FFNN), SVM | Accuracy, F-measure, MCC | FFNN Accuracy: 92.22% F-measure: 0.923 MCC: 0.835 SVM Accuracy: 92.40% F-measure: 0.924 MCC: 0.838 | Hájek and Olej (2013) |
Sentiment analysis of the bank’s customers | Bank reviews from mouthshut.com and myBankTracker.com | Ontology-driven sentiment analysis | Accuracy | 73.118% | Chaturvedi and Chopra (2014) |
Assessment of consumer financial complaints | Consumer complaints from the CFPB | SAS Contextual Analysis, SAS Visual Analytics, SAS Visual Statistics | Defining methods for text and visual analytics | – | Sabo (2017) |
Opinion mining analysis in the banking system | Social media text from Twitter and Facebook | Rough set theory combined with a decision system | Precision, recall, accuracy | Accuracy (best out of all): 79% | Sumathi and Sheela (2017) |
Analysis of bank reviews | Citibank reviews from Twitter, mouthshut.com, and myBankTracker.com | Opinion mining, sentiment analysis | Positive/negative | – | Gulaty (2016) |