Table 1 Computational complexity of several commonly used regression and classification techniques
From: A data value metric for quantifying information content and utility
Classifier | Type | Training | Prediction | Â |
|---|---|---|---|---|
Linear Regression | R | \({O(p^{2} n + p^{3} )}\) | \({O(p)}\) | (4) |
Decision Trees | C&R | \({O(n^{2} p)}\) | \({O(p)}\) | |
Random Forest | C | \({O(n^{2} pk_{{trees}} )}\) | \({O(pk_{{trees}} )}\) | |
Gradient Boosting | C&R | \({O(npk_{{trees}} )}\) | \({O(pk_{{trees}} )}\) | |
SVM | C&R | \({O(n^{2} p + n^{3} )}\) | \({O(m_{{sv}} p)}\) | |
k-Nearest Neighbors | C&R | varies | \({O(np)}\) | |
Neural Networks | C&R | varies | \({O(\sum _{i} o_{{l_{i} }} o_{{l_{{i + 1}} }} )}\) | |
Naive Bayes | C | \({O(np)}\) | \({O(p)}\) |