Forum

Like climate modeling, flood prediction also involves complex systems that are not linear and can behave in unexpected ways. Just as the video discusses the challenges of predicting exactly when major climate tipping points might occur, flooding prediction also has to grapple with the potential for sudden, nonlinear changes in water levels, and differing weather patterns. The similarities in modeling and simulation, even with inherent uncertainties, to try to anticipate future changes also exist. Flood predictions also rely heavily on sophisticated hydrological models and simulations that forecast things like river levels/storm surge/rainfall—there’s a level of adaptive or organic feedback loops that may be needed to continuously improve the system.

Professor Goodman's interview of Dan Gilbert relates to my final project in a few ways. My project involves the creation of my own model to predict the outcome of NBA playoff games. Part of my process includes deciding which factors to give weight to in my model. Things like home court advantage and team motivation are two psychological factors which are incredibly hard to model because we are uncertain about how each individual thinks about these inputs. I enjoyed listening to Dan talk about how humans perceive the difference between a "sure thing" and a "probable thing." For example, most people would say 2% likelihood is closer to 10% likelihood than it is to 0%. It is interesting how the mind will categorize information. To relate it back to my project, it is hard to judge a player's confidence when they might perceive a 2% chance of winning as the same "maybe/maybe not" scenario as a 50/50 chance of winning, thus leaving their motivation to win unchanged. More relevant information can be found in this sports psychology book. Vealey, Robin S. "Confidence in sport." Sport Psychology 1 (2009): 43-52.

As I am researching predictions in the Housing Market, I watched the interview with David Laibson on behavioral economics. Behavioral economics is deeply connected to predictions in the housing market because it explores how psychological, social, and emotional factors influence the economic decisions of individuals, including buying and selling homes. Additionally, Professor Laibson touched on his experience with machine learning and how this method can predict behavior economics. Having researched many prediction methods using some form of machine learning, it was interesting to hear about this in a more academic light. An additional source I have found is listed below. It discusses how the COVID-19 pandemic, with consumers having a total loss of control and freedom due to social distancing and other restrictions, has influenced how people shop for variety, such as in the wine market. This study uses machine learning and U.S. household data to find that people's variety-seeking behavior in wine purchases changed during 2020, initially decreasing with coupon use but later returning to normal and fluctuating with how frequently they shopped. This is just one example of how machine learning can predict behavioral economics and predict consumer preferences! Excited to further see how this affects prediction in the housing market.