Back to Article

Leveraging Machine Learning Techniques for Predictive Analysis in Merger and Acquisition (M&A)

Journal of Artificial Intelligence and Big Data | Vol 3, Issue 1

Table 1. Literature review summary for the mergers andacquisitions (M&A) analysis

Authors & Year	Focus/Objective	Methodology	Key Findings	Tools/Techniques	Limitations and Future Work
Yan, Nie, and Fan (2020)	Verify the mechanism of due diligence impact on M&A performance using data from Chinese A-share market (2011–2018).	Multivariate Regression Model	Higher diligence level in due diligence positively impacts M&A performance. Performance commitments affect the relationship by encouraging truthful disclosures and active integration.	Statistical analysis, Multivariate regression models	Limited to Chinese A-share market; future studies can explore cross-market comparisons and long-term impacts of performance commitments.
Shi and Wang (2020)	Develop an M&A matching method using DEA-based performance perspective that incorporates irrational behavior.	Data Envelopment Analysis (DEA) with a prospect value optimisation goal and cross-efficiency modelling	Proposed method effectively optimises two-sided M&A matching and can be used in third-party platforms for recommendation services.	DEA, Prospect value, Cross-efficiency modelling	Assumes availability of comprehensive data; future work could test the method's applicability in different industries and regions.
Li et al. (2021)	Investigate GNN application in M&A forecasting by quantifying statistical associations among firms, founders, and investors.	Graph Neural Networks (GNN), Text analysis, Feature engineering	Achieved an AUC of 0.952 and an 83% true positive rate in predicting M&A outcomes.	Graph Neural Networks, Text analysis, Statistical analysis	Focused on specific data sources, future studies could incorporate more diverse datasets and assess scalability in real-time predictions.
Lang (2021)	Evaluate technical and potential efficiency in virtual M&A using listed commercial banks in China as case studies.	Three-stage DEA and Bootstrap DEA	Virtual M&A is generally beneficial, but efficiency of M&A does not guarantee technical or potential efficiency post-M&A.	Three-stage DEA, Bootstrap DEA	Limited to banking sector; future research could apply the methodology to other industries and refine efficiency evaluation criteria.
Wang (2021)	Explore characteristics of enterprises and industries adopting M&A strategies in the U.S. (1978–2007) and their growth post-M&A.	Statistical analysis of historical M&A data using Stata	M&A strategic activities can lead to growth in scale, capital, and output value, varying by enterprise and industry.	Stata statistical analysis	Covers a specific timeframe; future work could explore post-2007 trends and technological influences on M&A strategies.
Venuti (2021)	Utilise Graph SAGE framework to predict M&A of enterprise companies.	Graph SAGE (Graph-based Machine Learning Framework)	Achieved 81.79% prediction accuracy, showcasing the effectiveness of graph-based ML for financial data science and M&A predictions.	Graph SAGE, Graph-based ML	Relies on graph structure assumptions; future research could test performance with more complex graph datasets and hybrid ML models.