App Designer for Classification

I have trained and tested a set of data with the RUSBoosted Tree ensemble based algorithm in classification learner.

However, I would like to apply the classification function in the App designer.

May I get some recommended ideas on how to do so?

Answer 1

[Image Analyst]

In the learner, export the model to a .mat file. Then in your app read in the mat file and call it like the mat file tells you to.

[Haw Jia Yong]

Thank you for your reply.

I had exported my model into .mat file and load into the app learner.

But how should I call the mat file to make the prediction?

I have 7 numeric data as input and 3 different output variables.

% Button pushed function: PredictButton

function PredictButtonPushed(app, event)

Classification=load('ClassificationLearnerSession1.mat');

Hydrogen=app.HydrogenEditField.Value;

Ethane=app.AcetyleneEditField.Value;

Dioxide=app.CarbonDioxideEditField.Value;

Monoxide=app.CarbonMonoxideEditField.Value;

Acetylene=app.AcetyleneEditField.Value;

Ethylene=app.EthyleneEditField.Value;

Methane=app.MethaneEditField.Value;

[Image Analyst]

You forgot to attach your exported mat file. When you read it in, it will tell you how to call it

Classification=load('ClassificationLearnerSession1.mat')  % No semicolon!!!

Then look in the command window. One of the fields of s should be a string telling you how to call it.

[Haw Jia Yong]

From the command window, I was asked to call by struct,

But how to allow my app.outputEditField to get the prediction output (conditionA, conditionB and conditionC) based on the input ?

Classification = struct('Hydrogen',Hydrogen, ...

'Ethane',Ethane, ...

'Dioxide',Dioxide, ...

'Monoxide',Monoxide, ...

'Acetylene',Acetylene, ...

'Ethylene',Ethylene, ...

'Methane',Methane)

[Image Analyst]

You forgot to attach the model in the .mat file again. And you also forgot to attach your .mlapp file. I'll check back later.

[Haw Jia Yong]

My current .mat file and .mlapp for your advice.