Statistical Learning Methods

Question 1

Bagging

Answer 1

A type of random forests that does the same thing but uses all features at each split because m = p

Question 2

Random Forests

Answer 2

using x number of bootstrapped training sets to construct x number of regression trees using a subset of features at each split then averaging the results

Question 3

Logistic Regression

Answer 3

Supervised machine learning algorithm widely used for binary classification tests

Question 4

Boosting

Answer 4

Boosting improves machine models’ predictive accuracy and performance by converting multiple weak learners into a single strong learning model

Question 5

Weak learners

Answer 5

Models with low prediction accuracy, similar to random guessing

Question 6

Strong learners

Answer 6

Models with high(er) prediction accuracy

Question 7

hierarchal clustering

Answer 7

Unsupervised learning method for clustering data observations. Clusters are built by measuring dissimilarity between data

Question 8

K means clustering

Answer 8

Divides observations into K number of clusters, each observation being assigned to the closest centroid of a cluster

Question 9

Ridge regression

Answer 9

Method of estimating the coefficients of multiple regression models in scenarios where predictors are highly correlated.

Shrinks coefficients but does not force it to 0, and therefore does not perform variable selection.

Question 10

Lasso

Answer 10

Regression analysis method that performs both variable selection and regularisation to enhance prediction accuracy and interpretability

Question 11

Regularisation

Answer 11

A set of methods for overfitted models to increase generalisability