Modeling the Judgments of Civil Cases of Support for the Elderly at the District Courts in Taiwan
摘要
The problem of legal judgment prediction (LJP) has attracted substantial attention from researchers of both the law and computer science communities in recent years. The majority of the previous LJP work is for criminal cases. We report our attempt to predict the judgments for the cases about the support for the elderly, which is an instance of the civil cases. We investigated the effects of choosing different design parameters in our decision models, which adopted traditional machine learning and deep learning concepts. The results of the empirical evaluations showed some encouraging results, but some others uncovered the challenges that remain to be tackled.