A New Semisupervised-Entropy Framework of Hyperspectral Image Classification Based on Random Forest

Table 1

The process of generate decision tree (using the given training data to produce a decision tree).

---

Algorithm: Generate decision tree(using the given training data to produce a decision tree)

---

Input: Training data set samples, with discrete value attributes, the collection of candidate attributes Attribute list;

Output: A decision Tree;

method: Create node N;

  If the samples are in the same Class C;

  returns N as a leaf node, with class C tag;

  If Attribute list is an empty then;

  returns N as the leaf node, marking the most common class in samples;//Majority Voting;

  Selecting the optimal classification attribute test attribute in Attribute list; Using information gain as attribute selection metric;

  The Mark node N is test attribute;

  for the known value AI in each test attribute; The Division of Samples

  It is grown from the node N with a condition of test attribute= ai;

  Set up SI as sample of test attribute=ai in samples; A partition;

  If si is an empty then;

  plus a leaf node, marked as the most common class in the samples;//majority vote

  else plus one by Generate decision tree (SI, Attribute_ List-test attribute) returns the node;

---