Скачать или смотреть SVM EasyTrainer - v1.0: Machine Learning - Get Full Regression Analysis with Support Vector Machine

Machine Learning Application Using MATLAB App Designer

SVM EasyTrainer - v1.0: Perform Full Regression Analysis with Support Vector Machine by Just One Click!

To download it, please visit its page on our official website:
https://matlab-apps.com/2023/08/29/sv...

In this video, we are going to show you a very beautiful app that makes regression process through support vector machine (SVM) just a fun! The app itself is called "SVM EasyTrainer", which emphasizes our allegation regarding its usefulness. It is highly user-friendly machine learning app that needs zero knowledge or experience of programming; specifically MATLAB programming codes. It can do all the basic and little advanced steps for you just by one or a few clicks!

By running the application, it can be clearly seen that the target vector can be selected from any column of the main table dedicated for the imported dataset, and the other columns will automatically be configured to represent the predictor(s) of the model. The entries can also be modified manually before starting the app if you would like that! The app accepts external datasets with any standard format; like .xls, .xlsx, .xlsb, .xlsm, and .csv spreadsheets.

From the control panel, the user can split the dataset to train and test subsets to reduce the overfitting phenomenon. The train subset is used to generate the model and the test subset is used to evaluate the performance of that model using new instances rather than those used in the training phase.

Furthermore, the user can select the evaluation metric (SAE, MSE, or RMSE) and whether if he/she want to reproduce the results; which is a very useful feature for comparison studies. After that, the user can select the kernel function (if it is polynomial, then the polynomial order can be set from 1 up to the 20th degree). Then, one of three possible solvers can be selected and the data itself can be standardize to enhance the sensitivity of the machine learning algorithm.

The plots can be fully customized as per user's needs. For example, for the actual and predicted responses, you can select the plot style, width, color, markers (with marker size, face color, and edge color), major and minor grid (color, width, line style, transparency), etc. Not just that, you can even export that plot and open it in an external figure with the same settings!

Add to that, the user can export all the tables for all the data sets and subsets, and save the SVM model with an m-file to run externally in the future using any other program or script.

Please note that this video presents the final product. You can download the project and unlock its full potential.