Automatic Machine Learning Toolbox

Version 1.07 (85.2 KB) by Umit Isikdag
Automatic Machine Learning (AutoML) Toolbox utilizes fitrauto/fitcauto functions for estimating with hyperparameter-optimized models.
Updated 24 May 2022

View License

This tools provides a GUI to MATLABs 'fitrauto' and 'fitcauto' functions to find hyperparameter optimized estimations for your classification and regression models.I hope you will find it useful.
Basic Usage Step by Step:
M: Use menu to load the data
1.Load your data from an MS Excel(xlsx) or Comma Seperated Values(csv) file.
2.Provide the test / train split ratio between 0 and 1. For a 100 rows data set 0.2 will split 80 rows for training and 20 rows for test.
M:Use menu to Split the data.
3.Select your problem type Classification/Regression.
4.Provide the column of Response Variable as integer value using the slider.You can select any column as the response variable.
5.Input MaxRuntime that you wish to run the algorithms in seconds.
6.If you like to use a subset of predictors,you can find feature importances for regression and classification models using FIMPC (for classification) and FIMPR(for regression) buttons.FIMPC uses fscchi2 function, FIMPR uses fsrftest function to calculate feature importances.Sometimes you can get inf values . Check MATLAB help pages of these functions.
7.If you want to use a subset of predictor variabales turn the UseSubset switch on, and put column numbers of selected predictors that you would like to use in training to the TrainWith box seperated by comma.You can use multiple predictor variables by inputting their column names : 2,4,5
8.You can select which learners to train with Learner Listboxes.
9.Actually you can also select multiple learners(ctrl+left click).This would be useful in comparison of several learner combinations.
M:Use the appropriate menu item to train your model.This can take time depending on your computer's configuration.Please wait with patience.Once the training is complete...
10.For Classification Models you can see the accuracy of the Test Set (TestAcc) between 0-1, and you can also view the Confusion Matrix (please switch the ShowCM button to On for this).
11.For the Regression Models you can get the RMSE of the Test set.
12.Please check your CPU loads and temprature in large datasets, i.e. datasets >1000 rows.
13.Please check these when training runs over 3 mins.Just to make sure your CPU is doing great :-)
14.Please be aware that due to their nature ensemble learners take more time to train than others.
15.Please be aware that the training time can be much longer than the Max Runtime you have entered in the related box.
16.Please start with short Max Runtimes such as 10 secs, and increase incrementally.
17.Please e-mail me for your questions I will be thankful if you can cite the tool in your academic papers.

Cite As

Umit Isikdag (2024). Automatic Machine Learning Toolbox (, MATLAB Central File Exchange. Retrieved .

MATLAB Release Compatibility
Created with R2021a
Compatible with R2021a and later releases
Platform Compatibility
Windows macOS Linux

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!
Version Published Release Notes

Definition change


Definition change


Name Change


Change in GUI


Summary Change


Name change and summary added