Many types of prediction models exist. Thus, we cannot
confirm if a certain prediction model is superior to others
in any application. Selecting an appropriate prediction
model for accidents in stadiums is critical [6]. Four
standards, namely, rationality, accuracy, stability, and
simplicity, are utilized to determine if a prediction model is
good or bad. Prediction models for accidents in stadiums
follow a development law with a specific environment;
thus, the prediction results of prediction models are
consistent with the development law [9]. The accuracy of
prediction models is also important. If prediction accuracy
does not satisfy the requirements, the prediction model is
unpractical regardless of how reasonable it is. In addition,
prediction models need have stability and simplicity and
must be able to accurately reflect the development and
changes in accident prediction over a long period of time.
Furthermore, they must be simple and easy to use, and
their computing time must be short. Only prediction
models that satisfy the four standards above can be
considered good and applicable to practical situations