인공신경망을 활용한 부동산 개발행위허가사건의 판결영향요인 분석

다국어 초록 (Multilingual Abstract)

The purpose of this study is to analyze the extent to which the causes of the administrative lawsuit cases related to the granting of the developmental act in real estate public law cases are significant by the computer. The method of analysis is the Artificial Neural Network Analysis of the variables that are significant in the outcome of the actual 319 judgments in all court of Korea. 23 factors such as area, land size, land title, ordinance violations, and environment inconsistency were selected as the independent variables. The dependent variable is whether or not to win a lawsuit. The Artificial Neural Network model is a method of machine learning based on the neural system of the brain of an organism
As a result of the analysis, Independent variables indicating normalization importance of 50% or more are as follows : ① whether applicant is corporation, ② whether land is in Seoul, Suwon, Chuncheon, Daejeon, Daegu, Ulsan, ③ land size, ④ land title, ⑤ square measure criterion, ⑥ civil complaints. Based on the results of this study, it is possible to refer to the results of research in the preparation of defense in the decision of whether or not to file a lawsuit related to grant of developmental act or the filed administrative litigation. Furthermore, it will increase understanding of the outcome of litigation.