Data Classification in Matlab

When you have data that is very clustered as one but of about 3 or 5 class labels, how do you make the dataset a well spread data set so that the classification models can clearly identify the different classes present.

Thank You in Advance

[Nicolas B.]

Could you provide any exmple to explain your question?

[DARLINGTON ETAJE]

See the picture attached

Answer 1

[Prabhan Purwar]

Hello,

Classification of clustered data can be achieved by projecting your data into a higher dimension, by making use of Linear Algebra techniques, which makes data more separable for clustering and classification purposes.

After projecting the data, make use of the following techniques, as illustrated in the following links:

• https://www.mathworks.com/help/stats/examples/classification.html(Basic Classification methods)
• https://www.mathworks.com/help/stats/classification-with-imbalanced-data.html(Classification with Imbalance Data)
• https://www.mathworks.com/help/deeplearning/ref/classify.html(Classification using Trained Convolution Network)
• https://www.mathworks.com/matlabcentral/fileexchange/28770-introduction-to-classification
• https://www.mathworks.com/help/stats/classificationlearner-app.html(Classificationlearner App)

Hope this helps

[DARLINGTON ETAJE]

Hi...thanks for your response. Please can you show me sample code for projecting your data into a higher dimension, by making use of Linear Algebra techniques