Carnapian Inductive Logic for Exponential Smoothing

Key Points

• This research examines the inductive logic underlying exponential smoothing and its implications for predictive probabilities.
• Analyzed qualitative invariance assumptions about conditional probabilities.
• Investigated connections to exchangeable processes and Bayesian inference.
• Explored relationships with kernel methods in machine learning.
• Discussed philosophical implications of probabilistic invariance and symmetries.
• Demonstrated that exponential smoothing can be derived from plausible qualitative assumptions.
• Highlighted the influence of recent observations on predictions.
• Established connections to various probabilistic frameworks and machine learning techniques.

Abstract