Dataset about discretionary damages of mental suffering in fatal car accident cases in Taiwan

We use the Taiwan Judicial Yuan Law and Regulations Retrieving System to search and collect the judgments from 2006 to 2020 in Taiwan Taichung District Court, and then we delete the irrelevant judgments such as judgments against the plaintiff (without assessing the mental suffering damages). Furthermore, this dataset only focuses on the judgments with one victim and one defendant, excluding the judgments with more than one victim or defendant.

Next, we take one plaintiff as one data unit, and we extract the legal factors from the judgments and take them as features in the dataset. Apart from the legal factors which judges considered while assessing the mental suffering damages, we add some extra features which we think important and relevant to the assessment of mental suffering dam- ages. For example, we add some features that might affect the discretionary damages, including the total number of plaintiffs who are victim’s parents in the same case, the total number of plaintiffs who are victim’s children in the same case, and the total number of plaintiffs in the same case.

Besides, we extract the discretionary damages from the judgments and take them as the target values (𝑦 vector) in the dataset.

Finally, there are some features in the dataset lacking the unified unit, so we need to determine the unified unit for those feature columns.

To sum up, the dataset contains 483 observations and 34 features