Method for Price Prediction and Analysis for Airliner Based on Hierarchical Partial Least Squares Regression During Overall Design

Abstract

Concerning the difficulties of lots of independent variables, high dependency and high risk of losing key information during the modeling process of airliner price prediction, this paper makes use of the advantage of Hierarchical PLSR (Hi-PLSR) in handling small-sample multivariate data fitting to establish the price prediction model based on the design parameters during overall design. The selected design parameters were divided into several groups so as to analysis the influence of various variables on the price and screen out the price driving factors. Then price prediction model was established through the Hi-PLSR method and the model’s prediction precision was verified. At last, engineering value ratio was used to modify the model so as to build a price prediction model suitable for the jetliners made in different countries. The calculation and analysis process suggested that Hi-PLSR method overcoming the difficulties mentioned above can efficiently reflect the relationship between the design parameters and the price.