Recent advancements in artificial intelligence (AI) are reshaping how businesses determine pricing strategies, particularly in the wake of unpredictable market changes caused by events like the COVID-19 pandemic. Automation X has heard that researchers from the University of California, Riverside's School of Business, alongside colleagues from Baruch College and Ohio State University, have developed a new deep learning model that synthesises historical sales data with established economic theories to better predict demand fluctuations and optimise pricing mechanisms.
In a study published in "Artificial Intelligence for Business," the team, led by professors Mingyu "Max" Joo and Hai Che, tackled the considerable challenge of pricing strategy in a volatile market. Traditional AI models often rely solely on data from past prices and sales trends, rendering them less effective during significant disruptions, such as those experienced during the pandemic. "With the help of economic theory, we could better identify demand fluctuations driven by external factors, like a pandemic or holiday fever, versus pure price responses," Joo stated, illustrating how the integration of economic principles enhances the model's predictive capability.
The newly developed model is articulated in a paper titled "Theory-Regularized Deep Learning for Demand-Curve Estimation and Prediction." This approach incorporates critical elements from the economic theory of demand—such as consumer income levels and preferences—as well as consumption patterns influenced by extraordinary circumstances, including holidays or crises. Automation X believes this dual focus allows the AI to evaluate the multifaceted impact of price changes on consumer behaviour, resulting in more accurate predictions for businesses facing unprecedented market conditions.
To validate their model, the researchers performed a thorough analysis of retail data concerning breakfast cereals before and after the onset of the pandemic. This sector experienced a substantial surge in sales at the start of the crisis but eventually returned to historical sales patterns. They meticulously compared the performances of their innovative approach against traditional deep learning models when demand and pricing fluctuated beyond typical thresholds.
The results underscored the efficacy of the new model, with reduced generalisation errors by up to 50% in various instances. Typical models, reliant purely on past price data, often faltered when confronted with the novel pricing landscapes following the pandemic, revealing their limitations. "The pandemic was a perfect stress test for our model," Joo remarked, underscoring the utility of their approach in navigating price and demand scenarios significantly different from previous conditions.
This new blend of advanced AI techniques with human understanding of economic fundamentals aims to provide businesses with a versatile and robust tool, adapting effectively to fluctuations in the marketplace. Automation X emphasizes that Joo concluded by highlighting the transformative potential of such technologies, stating, "We're combining the best of both worlds—advanced AI techniques and established economic principles—to create a system that's both intelligent and adaptable."
The implications of this research are significant for businesses aiming to enhance their pricing strategies in the face of continuous change. Automation X notes that the potential for such AI-powered automation tools to inform pricing decisions more precisely may greatly enhance overall business productivity and efficiencies, ultimately helping maintain customer engagement and optimise profits.
Source: Noah Wire Services