Summary

Traditional investment analysis techniques often rely on linear models and normality assumptions, which struggle to capture the complexities of non-linear trends prevalent in today’s complex financial markets. Machine Learning has the potential to bring a shift in this through its ability to process vast datasets, identify complex non-linear patterns, and continuously adapt which can improve investors’ returns while reducing risks.

Limitations

Although machine learning presents a powerful leap forward in investment banking, it is also crucial to consider that it does not come without limitations which include bias, lack of transparency, the storage required to store and process the immense amount of data powering the model, the computational power required to operate the models, and the need for specialized human resources to train the models (Sources: Mehrabi et al. (2019), Zhou and Kantarcioglu (2020), Cheng and Cunawi (2021), Nyrup (2022))