3.5. Procedure of STO algorithms
Based on the aforementioned descriptions of expert knowledge, data mining and HSA, the STO algorithms will be developed in what follows. The procedure given in Algorithm 3 illustrates the overall flowcharts in Fig. 3. In Algorithm 3, Steps 1–3 describe the model training phase, and Steps 4–9 describe the real-time train operation phase. Thus, by the fusion of expert knowledge and data mining algorithms, we obtain two STO algorithms, i.e., STOB and STOL, which are based on Bagging ( Algorithm 1) and LSBoost (Algorithm 2), respectively.