Data Mining Method for Identifying Biased or Misleading Future Outlook

Abstract

In this study, we introduce a data mining method to identify biased and/or misleading outlooks for future performance of various factors, such as income, corporate profits, production, countries’ GDP, etc. The method consists of several components. One very important component involves building a general model, where the dependent variable is a factor suspected of projecting an over-optimistic impression in some records. Explanatory variables in the model are viewed as representing the potential for the satisfactory performance of the dependent variable. The second component involves evaluating the potential for the individual records of interest (specific countries, corporations, production facilities, etc.), and allows us to identify possible gaps between the upbeat/optimistic projections into the future (of the dependent variable) versus low and/or declining potential. In other words, low and/or declining potential basically tells us that the optimistic future performance of the dependent variable is unattainable, and could also represent misleading or deceitful information. The important novelty of this study is the capability to identify a highly exaggerated outlook of future performance, by utilizing a soft regression tool and the concept of “performance potential”. The process is explained in detail, including the conditions for successful evaluations. Case studies to evaluate expected economic success are presented.