Agricultural Economics Department
First Advisor
Taro Mieno
Date of this Version
7-2019
Document Type
Thesis
Citation
A thesis presented to the faculty of the Graduate College at the University of Nebraska in partial fulfillment of requirements for the degree of Master of Science
Major: Agricultural Economics
Under the supervision of Professor Taro Mieno
Lincoln, Nebraska, July 2019
Abstract
Understanding the impact of different data processing methods on site-specific management recommendation is vital to famers and consultants. This thesis investigates how data processing methods affect variable selection in production function estimation and the consequent economically optimal input use recommendation map. We utilized LASSO, a regression analysis method for feature selection. Data are collected from two years of on-farm trials from the same field in Hamilton county, Nebraska. Results suggest that the feature selection process and final recommendation about seed and nitrogen rates can be very sensitive to the way we process and analyze data, and the effects can be very different in different years for the same field. It is very important for practitioners and researchers to raise awareness that different way of processing data may lead to different conclusions for field management and research studies.
Advisor: Taro Mieno
Comments
Copyright 2019, Zhengzheng Gao