Cash Flow Forecasting for Self-employed Workers: Fuzzy Inference Systems or Parametric Models?

Abstract Cash flow forecasting is an important task for any organization, but it becomes crucial for self-employed workers. In this paper, we model the cash flow of three real self-employed workers as a time series problem and compare the performance of conventional parametric methods against two types of fuzzy inference systems in terms of both prediction error and processing time. Our evaluation demonstrates that there is no winning model, but that each forecasting method’s performance depends on the characteristics of the cash flow data. However, experimental results suggest that parametric methods and Mamdani-type fuzzy inference systems outperform Takagi–Sugeno–Kang-type systems.