An Analysis of Changes Onion Yields in Korea using Panel Regression Analysis and Bayesian Network Model

This study examined the effects of meteorological and farm input cost factors on the onion yields in Korea by employing a panel regression analysis. Also, it investigated the variables’ interdependencies and their relations to the onion yields by using a Bayesian network model. We collected the panel data from 1991 to 2019 for our analysis. More specifically, we used the panel data of the regional meteorological factors by month (cumulative precipitation, cumulative sunshine duration, average relative humidity, average temperature), farm input cost factors of the onion (the cost of fertilizer and the cost of agricultural medicines), and the onion yield of three regions in Korea. We used STATA 14.0 and Hugin Expert for descriptive analysis, panel regression analysis, and the Bayesian network model. Our analysis can be summarized in two significant ways. First, we chose the fixed-effect model based on the Hausman test. The results based on the fixed-effect model confirmed that the average relative humidity (October, -), the cumulative precipitation (January, -; March, +; May, +), the cumulative sunshine duration (April, +), the average temperature (June, -), the cost of fertilizer(+) and the cost of agricultural medicines(+) were the significant variables of * 본 논문의 아이디어 구상에 도움을 주신 농촌진흥청 농업빅데이터일자리팀 조용빈 팀장님, 농촌진흥청과 네덜란드 바헤닝언대학 연구센터(WUR) 간 협력사업으로 추진된 ‘Bayesian Network Modelling & Big Data’ 훈련 프로그램에서 베이지안네트워크 분석 방법을 지도해 주신 Hans Marvin 박사님과 Yamine Bouzembrak 박사님, 그리고 본 논문을 심사해 주시고 유익한 조언을 해 주신 익명의 심사위원님들에게 깊은 감사를 드립니다. 더불어 논문 내용 중 있을 수 있는 오류는 저자들의 책임이며, 본 연구 결과는 저자들의 소속기관 공식 의견이 아닌 개인 의견임을 밝힙니다. ** 농촌진흥청 농업빅데이터일자리팀 박사후연구원. 교신저자. e-mail: silee79@korea.kr *** 농촌진흥청 농업빅데이터일자리팀 농업연구관. **** 농촌진흥청 농업빅데이터일자리팀 농업연구사.