Home

Datascience

Correlation Testing Geo-Timeseries Linear Regressions Machine Learning Data Science Data Visualization

Don't Look! I'm changing!

URL Copied

IntroductionProblem ClassesRegression-Type Problems.Classification-Type Problems.Tree-Based SolutionsModel Types By Statistical RoleGenerativeDiscriminativeDescriptiveLearning Model ParadigmsSupervised:UnsupervisedSemi-SupervisedSelf-SupervisedReinforcement Learning

Introduction

A learning model is a parameterized function whose parameters are learned from data. When data is processed through a learning model it's outputs are often approximate relationships or make predictions.

Problem Classes

There are numerous algorithms for predicting continuous variables or categorical variables from a set of continuous predictors and/or categorical factor effects.

Regression-type problems.

Regression-type problems are generally those where we attempt to predict the values of a continuous variable from one or more continuous and/or categorical predictor variables.

Classification-type problems.

Classification-type problems are generally those where we attempt to predict values of a categorical dependent variable (class, group membership, etc.) from one or more continuous and/or categorical predictor variables. There are a number of methods for analyzing classification-type problems and to compute predicted classifications, either from simple continuous predictors (e.g., binomial or multinomial logit regression in GLZ), from categorical predictors (e.g., Log-Linear analysis of multi-way frequency tables), or both (e.g., via ANCOVA-like designs in GLZ or GDA).

Tree-Based Solutions

Tree methods are nonparametric and nonlinear + Simplicity of results. -Specify Criteria for Predictive Accuracy, Selecting Splits, When to Stop Splitting.


Model Types By Statistical Role

Generative

Discriminative

Descriptive

Learning Model Paradigms

Supervised:

labeled inputs; regression and classification.

Unsupervised

latent structure; clustering, density estimation, manifold learning.

Semi-supervised

joint usage of labeled + unlabeled samples.

Self-supervised

surrogate labels from transformations; pretraining regimes.

Reinforcement learning

sequential decision processes maximizing cumulative reward.