Behavorial Economics + Machine Learning

Behavorial economics is the study of psychological and social factors that influence the economy. It is based on the idea that as humans we are naturally accustomed to patterned behaviors, that our actions are repetitive. As this field of study has begun to advance, so has the capabilities of machine learning. Because of the predictive capabilities of machine learning, it could be extremely beneficial in reading the economy and how the market moves based on emotional sentiment. Which is why today, I will be talking about the potential capabilities of machine learning in the lens of behavorial economics.

There are many ways that machine learning could enhance the study of behavorial economics. The most obvious benefit being that it would be able to predict consumer behavior by analyzing past trends as well as current data (news outlets and social media) that help indicate what the current public sentiment is. Although machine learning approaches in the field of behavorial economics hasn’t been heavily talked about, I was able to find an article written by Ahmad Tanehkar about the applications of machine learning in behavorial economcs that he wrote in 2019. In the article, one model that he highlights the importance of is Random Forest. As I covered in a previous post, Random forest is a model for prediction that generates data through decision “trees”. For example, if it was trying to determine what sickness a person has, it would generate “random trees” that look at all the aspects of the sickness. These trees help clarify whether the patient has a headache or stomach pain or fever and then combine all the data once it has acquired it to make a final decision. In the case of behavorial economics, Random Forest would be useful because it can handle and assess all the variables necessary in predicting an economic outcome. Things like psychological traits and common behavior, all things that have an affect on the economy, can all be assessed. Additionally, random forest also has the capability to handle missing data which can also make it useful in the context of behavorial economics. More specifically, if someone were to encounter a dataset with missing values, random forest would be able to construct more “trees” by using the subsets of the data it was given. This would be extremely useful for a situation where, for example, a survey was given out on consumer spending and some questions were not answered.

In conclusion, the integration of machine learning into behavioral economics holds significant promise for enhancing our understanding and prediction of economic trends. By leveraging models like Random Forest, which can analyze complex and incomplete data, machine learning offers the potential to more accurately forecast consumer behavior based on psychological traits and social patterns. As this approach continues to evolve, it could provide powerful tools for economists to better interpret market movements and the underlying emotional sentiments that drive them.