**3. Materials and Methods**

This study proposes a safety behavior classification framework that combines statistical analysis methods and machine learning algorithms. As shown in Figure 2, the framework has three steps, i.e., data collection and preprocessing, modeling and algorithm implementation, and optimal model acquisition. The data is processed automatically by the proposed combinative strategies. The proposed methods are described in detail as follows.

**Figure 2.** The safety behavior classification framework. First, users need to determine variables and indicators, and complete necessary preprocessing after data collection. Second, each trial has a unique code, and 64 models in total are trained and tuned automatically by specific methods mentioned in their codes. Last, the performance of the 64 models is output, and the model with maximum scores stands out as the optimal model. Meanwhile, users can also observe the results of feature selection to guide the analysis of the important factors of one risk behavior or the average important factors of certain risk behaviors.
