Student Grade Prediction System Using Improved Random Forest Algorithm.​

Saiman Dahal. Washington State University

Improved Random forest: It is the combined process of bootstrapping without replacement, random split node and aggregation of the output.

Algorithm:

Pick a random split.​

Input: subset S and an attribute a​

Output: a split​

K number of attributes randomly selected at each node. (a1,...ak)​

A random cut-point ac selected uniformly in [amax , amin]​

Return the split [a < ac].​

If all attributes are in the subset S, then stop split.​

Return a split such that max(Score(S))​