Need help about multiclass classifier for Matlab

i need 2 matlab function . These functions will be multiclass classifier . I have 26 different classes . each class has 16 different values . i have 20k rows data at total . each row has class label (integer) and 16开发者_如何转开发 different values as i said . first 16 k rows of the data will be used for training . other remaining 4k rows will be used for estimating classification performance . so i will use this 2 function for classify them and compare their performance .

thanks a lot for every kind of contribution.

---

Statistics Toolbox provides a variety of different multiclass classifiers including

Bagged Decision Trees Naive Bayes Classifiers Linear Discriminant Analysis

You can access the product user guide at http://www.mathworks.com/help/toolbox/stats/f15360dfi14.html

Alternatively, you might want to consult a recorded webinar titled "Computational Statistics: Getting Started with Classification using MATLAB"

---

I'm not sure what the "2 functions" are, but it sounds like your basic need is for a classification routine.

If you have the Statistics Toolbox, try the classify() or classregtree() functions (there are other appropriate functions, but these are probably the easiest to apply).

If you don't have the Statistics Toolbox, try searching online, especially at MATLAB Central. I will immodestly suggest my own LDA() function, which performs linear discriminant analysis and can handle multiple classes, described here:

Data Mining in MATLAB: Linear Discriminant Analysis (LDA)