A Nicholas Center team of MBA and BBA students completed our first-ever machine learning consulting project this semester. Congrats to the team – James Kardatzke, Mason Bourne, Caleb White, Mary Roberts, Dan Miller and Thomas Dircks!
The team developed everything from scratch – data selection, data collection, developing the code in Python, analysis of results, etc.
The team produced:
- a research report (Research Report Predicting Mergers and Acquisitions With Machine Learning),
- a detailed process manual that explains all of the key steps in the analysis (ProcessManual_MachineLearningProject_NicholasCenter), and
- their code is publicly available on github (https://github.com/UW-Nicholas-Center/Adage)
A few key highlights from this analysis:
- Problem statement – which public companies will be the target of an M&A acquisition in the next 12 months?
- Collected and analyzed over 1 billion data points (254 variables on a quarterly basis from 1990-2019)
- Utilized neutral network, random forest and ensemble models for prediction
- The results were 6x more accurate than the base rate
- We highlight and perform fundamental analysis on the 10 companies that our model predicts will be the target of M&A in the next 12 months
It is important to note that this is the Nicholas Center’s first experiment using machine learning to answer a finance question. While the results are not sufficient for practical use as a standalone tool, this provide an excellent learning opportunity and the analysis provides a foundation that can be improved upon down the road.