Community Profile

# Ilya

### MathWorks

228 total contributions since 2011

View details...

Contributions in
View by

What formula is used to calculate perplexity in fitlda?
If you are asking about the 2nd output from the logp method, document log-probabilities are estimated using the Mean-Field Appro...

alrededor de 1 año ago | 1

| accepted

Matlab: Error using classreg.learning.FitTemplate/fit with hyperparameter optimization of SVM
You are passing ClassNames to fitcecoc - are your ClassNames a subset of all class names you have in yTrain? Train one ECOC m...

más de 1 año ago | 0

| accepted

How are the automatic values of hyper-parameters in Matlab Regression Learner determined?
If you type edit classreg.learning.svmutils.optimalKernelScale in your MATLAB session and hit Return, the editor will brin...

casi 2 años ago | 0

Perform Naive-Bayes classification(fitcnb) with non-zero off-diagonal covariance matrix
To estimate covariance per class, use |fitcdiscr| with discriminant type 'quadratic'.

más de 2 años ago | 0

Lack of fit with fitrlinear on multivariate data (version 2016a and later)
Your test set has floor(74*0.05)=3 observations. You can't measure error of any model on such a tiny test set.

más de 2 años ago | 0

To calculate mahalanobis distance when the number of observations are less than the dimension
For classification, use regularized discriminant or pseudo discriminant. Both options are supported in |fitcdiscr|. Regularizati...

más de 2 años ago | 0

TreeBagger - Random forest
Take one observation and compute prediction for that observation. Then replace some predictor in that observation with NaN and r...

casi 3 años ago | 0

kmeans appear to miss obvious clusters
Do this (assuming there are no nan's in X): [cidx3,cmeans2] = kmeans(zscore(X),3,'dist','cosine','display','iter'); Did ...

casi 3 años ago | 1

| accepted

Help assess my Random Forest work and work on feature selection
First, you would increase your chance of getting a useful reply if you simplified the problem. Your code and your question are r...

casi 3 años ago | 1

Understanding the equations behind the 'logistic' learner when using fitclinear
<https://www.mathworks.com/help/stats/classificationlinear.predict.html#bu4z0pc-6 predict help>: If the linear classification...

casi 3 años ago | 0

| accepted

Nonlinear regression with categorical predictor?
Unless I misunderstood your dot notation, the problem is ill-defined. It has an infinite number of solutions. Rewrite it in this...

alrededor de 3 años ago | 2

What does it mean for a tree in a TreeBagger ensemble to have to have >80% error? What is the best way to reduce error?
It's hard to identify the source of discrepancy without understanding what the package at that link does and how you used it. Ho...

alrededor de 3 años ago | 0

| accepted

помогите с загрузкой Matlab
Алексей, на этом форуме редко кто говорит по-русски. По вопросам установки лучше всего обращаться в техническую поддержку в стра...

alrededor de 3 años ago | 1

How to use svm in Matlab for my binary feature vector.
You most certainly do not need as many samples as you have features. Statements like "you need at least 6 times the number of ca...

alrededor de 3 años ago | 0

Why is SVM performance with small random datasets so high?
Let me make sure I got your procedure right. You apply M models to a dataset and measure their accuracies by cross-validation. E...

alrededor de 3 años ago | 0

| accepted

Classifier cross validation on grouped observations with different class ratio's
I haven't understood what you mean by "performing cross validation on 'grouped' observations. Where the patient ID would corresp...

más de 3 años ago | 0

Why is SVM performance with small random datasets so high?
You have 12 observations. For each observation, the probability of correct classification is 0.5. What is the probability of cla...

más de 3 años ago | 0

Feature selection to perform classification using Multinomial Logistic Regression
You describe a procedure for selecting a set of features at fixed hyperparameter values. You do not say what you do, if anything...

más de 3 años ago | 0

Feature selection to perform classification using Multinomial Logistic Regression
You should not use a linear model for feature selection and a nonlinear model for classification on the selected features. If yo...

más de 3 años ago | 0

| accepted

How to solve the “out of memory” problem in Logistic Regression achieved by “glmfit”
Logistic regression on tall (out-of-memory) arrays is supported in 16b through the |fitglm| function.

más de 3 años ago | 1

fitcsvm with identical variables gives different result on different machines
My guess is that gradients for two or more observations become equal within floating-point accuracy during optimization. The sol...

más de 3 años ago | 1

| accepted

How to get the mean of ROC curves using Matlab?
Use |perfcurve|. Take a look at this piece of <http://www.mathworks.com/help/stats/perfcurve.html#bupt4p4-3 documentation>. Pass...

más de 3 años ago | 0

statistical significance at the 95% confidence level
The reference is T. Dietterich. Approximate statistical tests for comparing supervised classification learning algorithms....

casi 4 años ago | 2

| accepted

How can I change the properties of a classification model template?
modelTemplate.ModelParams.BoxConstraint = 100; This is undocumented and can change in a future release.

casi 4 años ago | 2

| accepted

definition of score when using "predict" on trained adaBoostM1
Going to MATLAB online doc, typing 'AdaBoost" in the search box and then selecting the 3rd match brings me to this page: http...

casi 4 años ago | 0

Matlab SVM linear binary classification failure
This is a consequence of the data being poorly scaled. Do std(m3) and observe that the standard deviations of the two predictors...

alrededor de 4 años ago | 0

| accepted

perfcurve and ROC curve
Q must be classification scores. What you put in Q sounds more like what perfcurve should return as output. Take a classifier fr...

más de 4 años ago | 0

One standard error rule, classification in Matlab
The purpose is to simplify the tree without losing too much accuracy. One standard error is a heuristic rule. If one number is w...

más de 4 años ago | 0

| accepted

How can I make a decision stump using a decision tree ?
Use fitctree(X,Y,'minparent',size(X,1),'prune','off','mergeleaves','off')

más de 4 años ago | 1

| accepted