Synthetic Data and Generative AI

Chapter 9: New interpolation methods for synthetization and prediction

Abstract

I describe little-known original interpolation methods with applications to real-life data sets. These simple techniques are easy to implement and can be used for regression or prediction. They offer an alternative to model-based statistical methods. Applications include interpolating ocean tides at Dublin, predicting temperatures in the Chicago area with geospatial data, and a problem in astronomy, namely planet alignments and frequency of these events. In one example, the 5-min data can be replaced by 80-min measurements, with the 5-min increments reconstructed via interpolation, without noticeable loss. Thus, my algorithm can be used for data compression.

The first ...