resubLoss

Resubstitution regression loss

Syntax

L = resubLoss(Mdl)

L = resubLoss(Mdl,Name=Value)

Description

L = resubLoss(Mdl) returns the regression loss by resubstitution (L), or the in-sample regression loss, for the trained regression model Mdl using the training data stored in Mdl.X and the corresponding responses stored in Mdl.Y.

The interpretation of L depends on the loss function (LossFun) and weighting scheme (Mdl.W). In general, better models yield smaller loss values. The default LossFun value is "mse" (mean squared error).

L = resubLoss(Mdl,Name=Value) specifies additional options using one or more name-value arguments. For example, IncludeInteractions=false specifies to exclude interaction terms from a generalized additive model Mdl.

Examples

Resubstitution Loss

Train a generalized additive model (GAM), then calculate the resubstitution loss using the mean squared error (MSE).

Load the patients data set.

load patients

Create a table that contains the predictor variables (Age, Diastolic, Smoker, Weight, Gender, SelfAssessedHealthStatus) and the response variable (Systolic).

tbl = table(Age,Diastolic,Smoker,Weight,Gender,SelfAssessedHealthStatus,Systolic);

Train a univariate GAM that contains the linear terms for the predictors in tbl.

Mdl = fitrgam(tbl,"Systolic")

Mdl = 
  RegressionGAM
            PredictorNames: {'Age'  'Diastolic'  'Smoker'  'Weight'  'Gender'  'SelfAssessedHealthStatus'}
              ResponseName: 'Systolic'
     CategoricalPredictors: [3 5 6]
         ResponseTransform: 'none'
                 Intercept: 122.7800
    IsStandardDeviationFit: 0
           NumObservations: 100

Mdl is a RegressionGAM model object.

Calculate the resubstitution loss using the mean squared error (MSE).

L = resubLoss(Mdl)

L = 
4.1957

Compute Custom Resubstitution Loss

Load the sample data and store in a table.

load fisheriris
tbl = table(meas(:,1),meas(:,2),meas(:,3),meas(:,4),species,...
'VariableNames',{'meas1','meas2','meas3','meas4','species'});

Fit a GPR model using the first measurement as the response and the other variables as the predictors.

mdl = fitrgp(tbl,'meas1');

Predict the responses using the trained model.

ypred = predict(mdl,tbl);

Compute the mean absolute error.

n = height(tbl);
y = tbl.meas1;
fun = @(y,ypred,w) sum(abs(y-ypred))/n;
L = resubLoss(mdl,'lossfun',fun)

L = 
0.2345

Compare GAMs by Examining Regression Loss

Train a generalized additive model (GAM) that contains both linear and interaction terms for predictors, and estimate the regression loss (mean squared error, MSE) with and without interaction terms for the training data and test data. Specify whether to include interaction terms when estimating the regression loss.

Load the carbig data set, which contains measurements of cars made in the 1970s and early 1980s.

load carbig

Specify Acceleration, Displacement, Horsepower, and Weight as the predictor variables (X) and MPG as the response variable (Y).

X = [Acceleration,Displacement,Horsepower,Weight];
Y = MPG;

Partition the data set into two sets: one containing training data, and the other containing new, unobserved test data. Reserve 10 observations for the new test data set.

rng('default') % For reproducibility
n = size(X,1);
newInds = randsample(n,10);
inds = ~ismember(1:n,newInds);
XNew = X(newInds,:);
YNew = Y(newInds);

Train a generalized additive model that contains all the available linear and interaction terms in X.

Mdl = fitrgam(X(inds,:),Y(inds),'Interactions','all');

Mdl is a RegressionGAM model object.

Compute the resubstitution MSEs (that is, the in-sample MSEs) both with and without interaction terms in Mdl. To exclude interaction terms, specify 'IncludeInteractions',false.

resubl = resubLoss(Mdl)

resubl = 
0.0292

resubl_nointeraction = resubLoss(Mdl,'IncludeInteractions',false)

resubl_nointeraction = 
4.7330

Compute the regression MSEs both with and without interaction terms for the test data set. Use a memory-efficient model object for the computation.

CMdl = compact(Mdl);

CMdl is a CompactRegressionGAM model object.

l = loss(CMdl,XNew,YNew)

l = 
12.8604

l_nointeraction = loss(CMdl,XNew,YNew,'IncludeInteractions',false)

l_nointeraction = 
15.6741

Including interaction terms achieves a smaller error for the training data set and test data set.

Input Arguments

Mdl — Regression machine learning model

full regression model object

Regression machine learning model, specified as a full regression model object, as given in the following table of supported models.

Model	Regression Model Object
Gaussian process regression model	RegressionGP
Generalized additive model (GAM)	RegressionGAM
Neural network model	RegressionNeuralNetwork

Name-Value Arguments

Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

Before R2021a, use commas to separate each name and value, and enclose Name in quotes.

Example: resubLoss(Mdl,IncludeInteractions=false) excludes interaction terms from a generalized additive model Mdl.

IncludeInteractions — Flag to include interaction terms

true | false

Flag to include interaction terms of the model, specified as true or false. This argument is valid only for a generalized additive model. That is, you can specify this argument only when Mdl is RegressionGAM.

The default value is true if Mdl contains interaction terms. The value must be false if the model does not contain interaction terms.

Example: IncludeInteractions=false

Data Types: logical

LossFun — Loss function

"mse" (default) | function handle

Loss function, specified as "mse" or a function handle.

• "mse" — Weighted mean squared error.

• Function handle — To specify a custom loss function, use a function handle. The function must have this form:

  lossval = lossfun(Y,YFit,W)

  • The output argument lossval is a floating-point scalar.

  • You specify the function name (lossfun).

  • If Mdl is a model with one response variable, then Y is a length-n numeric vector of observed responses, where n is the number of observations in Tbl or X. If Mdl is a model with multiple response variables, then Y is an n-by-k numeric matrix of observed responses, where k is the number of response variables.

  • YFit is a length-n numeric vector or an n-by-k numeric matrix of corresponding predicted responses. The size of YFit must match the size of Y.

  • W is an n-by-1 numeric vector of observation weights.

Example: LossFun=@lossfun

Data Types: char | string | function_handle

OutputType — Type of output loss

"average" (default) | "per-response"

Since R2024b

Type of output loss, specified as "average" or "per-response". This argument is valid only for a neural network model. That is, you can specify this argument only when Mdl is a RegressionNeuralNetwork object.

Value	Description
"average"	resubLoss averages the loss values across all response variables and returns a scalar value.
"per-response"	resubLoss returns a vector, where each element is the loss for one response variable.

Example: OutputType="per-response"

Data Types: char | string

PredictionForMissingValue — Predicted response value to use for observations with missing predictor values

"median" | "mean" | "omitted" | numeric scalar

Since R2023b

Predicted response value to use for observations with missing predictor values, specified as "median", "mean", "omitted", or a numeric scalar. This argument is valid only for a Gaussian process regression or neural network model. That is, you can specify this argument only when Mdl is a RegressionGP or RegressionNeuralNetwork object.

Value	Description
"median"	resubLoss uses the median of the observed response values in the training data as the predicted response value for observations with missing predictor values. This value is the default when Mdl is a RegressionGP or RegressionNeuralNetwork object.
"mean"	resubLoss uses the mean of the observed response values in the training data as the predicted response value for observations with missing predictor values.
"omitted"	resubLoss excludes observations with missing predictor values from the loss computation.
Numeric scalar	resubLoss uses this value as the predicted response value for observations with missing predictor values.

If an observation is missing an observed response value or an observation weight, then resubLoss does not use the observation in the loss computation.

Example: PredictionForMissingValue="omitted"

Data Types: single | double | char | string

StandardizeResponses — Flag to standardize response data

false or 0 (default) | true or 1

Since R2024b

Flag to standardize the response data before computing the loss, specified as a numeric or logical 0 (false) or 1 (true). This argument is valid only for a neural network model. That is, you can specify this argument only when Mdl is a RegressionNeuralNetwork object.

If you set StandardizeResponses to true, then the software centers and scales each response variable in Mdl.Y by the corresponding column mean and standard deviation. Specify StandardizeResponses as true when you have multiple response variables with very different scales and OutputType is "average". Do not standardize the response data when you have only one response variable.

Example: StandardizeResponses=true

Data Types: single | double | logical

Output Arguments

L — Regression loss

numeric scalar | numeric vector

Regression loss, returned as a numeric scalar or vector. The type of regression loss depends on LossFun.

When Mdl is a model with one response variable, L is a numeric scalar. When Mdl is a model with multiple response variables, the size and interpretation of L depend on OutputType.

More About

Weighted Mean Squared Error

The weighted mean squared error measures the predictive inaccuracy of regression models. When you compare the same type of loss among many models, a lower error indicates a better predictive model.

The weighted mean squared error is calculated as follows:

$$\text{mse}=\frac{{\displaystyle \sum _{j=1}^n{w}_j{\left(f\left(x_j\right)-y_j\right)}^2}}{{\displaystyle \sum _{j=1}^n{w}_j}},$$

where:

• n is the number of rows of data.

• x_j is the jth row of data.

• y_j is the true response to x_j.

• f(x_j) is the response prediction of the model Mdl to x_j.

• w is the vector of observation weights.

Algorithms

resubLoss computes the regression loss according to the corresponding loss function of the object (Mdl). For a model-specific description, see the loss function reference pages in the following table.

Model	Regression Model Object (Mdl)	loss Object Function
Gaussian process regression model	RegressionGP	loss
Generalized additive model	RegressionGAM	loss
Neural network model	RegressionNeuralNetwork	loss

Alternative Functionality

To compute the response loss for new predictor data, use the corresponding loss function of the object (Mdl).

Extended Capabilities

GPU Arrays

Accelerate code by running on a graphics processing unit (GPU) using Parallel Computing Toolbox™. (since R2024b)

This function fully supports GPU arrays for RegressionNeuralNetwork model objects. For more information, see Run MATLAB Functions on a GPU (Parallel Computing Toolbox).

Version History

Introduced in R2015b

R2024b: Compute loss for neural network regression model trained with multiple response variables

You can create a neural network regression model with multiple response variables by using the fitrnet function. Regardless of the number of response variables, the function returns a RegressionNeuralNetwork object. You can use the resubLoss object function to compute the resubstitution regression loss.

In the call to resubLoss, you can specify to return the average loss or the loss for each response variable by using the OutputType name-value argument. You can also specify whether to standardize the response data before computing the loss by using the StandardizeResponses name-value argument.

R2024b: Specify GPU arrays (requires Parallel Computing Toolbox)

resubLoss fully supports GPU arrays for RegressionNeuralNetwork model objects.

R2023b: Specify predicted response value to use for observations with missing predictor values

Starting in R2023b, when you predict or compute the loss, some regression models allow you to specify the predicted response value for observations with missing predictor values. Specify the PredictionForMissingValue name-value argument to use a numeric scalar, the training set median, or the training set mean as the predicted value. When computing the loss, you can also specify to omit observations with missing predictor values.

This table lists the object functions that support the PredictionForMissingValue name-value argument. By default, the functions use the training set median as the predicted response value for observations with missing predictor values.

Model Type	Model Objects	Object Functions
Gaussian process regression (GPR) model	RegressionGP, CompactRegressionGP	loss, predict, resubLoss, resubPredict
	RegressionPartitionedGP	kfoldLoss, kfoldPredict
Gaussian kernel regression model	RegressionKernel	loss, predict
	RegressionPartitionedKernel	kfoldLoss, kfoldPredict
Linear regression model	RegressionLinear	loss, predict
	RegressionPartitionedLinear	kfoldLoss, kfoldPredict
Neural network regression model	RegressionNeuralNetwork, CompactRegressionNeuralNetwork	loss, predict, resubLoss, resubPredict
	RegressionPartitionedNeuralNetwork	kfoldLoss, kfoldPredict
Support vector machine (SVM) regression model	RegressionSVM, CompactRegressionSVM	loss, predict, resubLoss, resubPredict
	RegressionPartitionedSVM	kfoldLoss, kfoldPredict

In previous releases, the regression model loss and predict functions listed above used NaN predicted response values for observations with missing predictor values. The software omitted observations with missing predictor values from the resubstitution ("resub") and cross-validation ("kfold") computations for prediction and loss.