Comparison of Interpolation and Regression Methods for Predicting Oil Royalties in Peru

Authors

Adriano, Leonardo

Tapia, Antonia

Arévalo, Clemente

Sinchitullo, Joseph

Abstract

The pandemic has dealt a heavy blow to the hydrocarbon industry, and with it, the collection of oil royalties has also been affected. For this reason, this research aims to predict the collection of oil royalties for the 4 months after the available history, its interpretation, and comparison between models applied in the study. For this, interpolation and polynomial regression methods were used based on the history of royalties, applying a cross-sectional methodology due to the simplicity in the interpretation of data and statistical graphs. It is shown that it is not feasible to consider all the points as interpolating nodes, in addition, small variations of the data in the graph can alter the prediction curve, causing them to have an increasing or decreasing behavior. Finally, a segmentation analysis was performed on the historical curve of royalties to test the existence of patterns or periods with a fixed trend with the help of the prediction curves.