[Retracted] Rapid Identification of Tobacco Mildew Based on Random Forest Algorithm

Table 1

Advantages and disadvantages of decision tree.

---

Advantages

1. There is no need to introduce a priori assumption.

2. The decision tree is relatively simple to understand, its logical thinking is easier for people to understand, and it is easy to realize in practical application.

3. It has good stability for outliers and noise. The decision tree does not classify the data according to the specific value, and the outliers in the data have little impact on the whole result.

---

Disadvantages

1. In the process of top-down recursive construction, nodes store less and less information, and too little information will cause data fragmentation.

2. In the process of modeling, the unstable splitting of leaf nodes will cause overfitting.

---