A hypothesis space is the set of candidate model classes available for solving a particular prediction or pattern-finding problem. Each class groups formulas and algorithms with shared characteristics and a common purpose. section 9.1.1
The problem usually narrows the choice: regression targets continuous outcomes, classification targets categorical outcomes, and other classes address grouping, time-to-event, time-series, or dimension-reduction tasks. section 9.1.1
Choosing a model class
Predicting store spending calls for a continuous-outcome approach, identifying whether students pass calls for classification, and forecasting sales over nine months calls for a time-series approach. section 9.1.1
Table 9.1 Classes of models, their purposes, and examples
|
Class
|
Purpose
|
Example
|
|---|---|---|
|
Regression
|
Predict outcomes represented with continuous data
|
Predicting student test scores
|
|
Classification
|
Predict outcomes represented with categorical data
|
Predict which students will graduate on time
|
|
Clustering
|
Identify underlying patterns and natural “groups” in the data
|
Find clusters of student performance across classes
|
|
Dimension reduction
|
Identify shared variance among predictors to simplify model input
|
Find 2 underlying components to represent scores in 12 classes
|
|
Survival analysis
|
Predict the time until an event occurs
|
Predict the time to withdraw from an intensive graduate curriculum
|
Compare model classes by their purpose and representative application, from regression and classification to clustering, dimension reduction, and survival analysis. section 9.1.1
