search_delta_n_samples

rpart.tree. A Decision tree generated by rpart package.

rpart.tree

int. The # of samples to be used as initial value to search the minimum necessary to ensure learning.

n_samples

float. The initial value of delta (need to be greater than epsilon)

delta

float. The epsilon to be used in the delta calculation.

epsilon

Search the # of samples to ensure learning given an epsilon.

Learning, in Machine Learning (ML) area, is one of the most important steps in the construction of algorithms that seek to predict a certain task, whether this is the classification of objects, the forecast of demand for a specific product or even the diagnosis of malignant diseases. In ML, we can study supervised (which have a label, e.g., a class) and unsupervised algorithms, used for tasks such as pattern detection, grouping, among others that do not depend directly on a label. Knowing this, the present work aims to carry out the study of different supervised learning algorithms, in this case, the classification algorithms, more specifically Decision Trees, to carry out an analytical study about the steps that make up the learning process of the algorithm, exploring concepts of the statistical learning theory (SLT) that provide tools for studies and allow to prove issues such as the guarantee of learning of a certain algorithm. Reference: Rodrigo Fernandes de Mello, Chaitanya Manapragada, Albert Bifet: "Measuring the Shattering coefficient of Decision Tree models". Expert Syst. Appl. 137: 443-452 (2019)<DOI:10.1016/j.eswa.2019.07.012>.

search_delta_n_samples: Search the # of samples to ensure learning given an epsilon.

Description

Usage

Arguments

Value