Assessing financial advice using machine learning

Abstract: The fields of artificial intelligence and machine learning are rapidly transforming the global technological landscape. Many enterprises within the financial sector are increasingly pursuing such projects to replace or enhance existing systems to be part of the new technological frontier. ABC Bank is a financial institution that aims to introduce a level of screening of advisory-texts in order to enhance quality and to improve resource efficiency within the firm. Through development of an implementation using the models Random Forest, Multi-Layer Perceptron and Support Vector Machine, this project served as a pilot study, an enabler, for further machine learning solutions to be developed within the bank. Whereas the developed model performed better than randomly guessing the most likely class-label for an advisory text, the scarce amount of data currently accessible prevents making claims whether the implementation is eligible for production-stage or not. The project highlights some barriers for introducing machine learning on a larger scale and advice on how to proceed is presented.