MATLAB: Using exported classifer model

classification learnermachine learning

Hi everyone,
I wonder if anyone out there can assist. I am learning to use the classification learner app in Matlab2015a. I generated some 'toy' data with 2 columns, choosing column 1 as predictor and column 2 as the response. 1 trained the CosineKNN classifer on this toy data and exported my model. I would now like to make some predictions using some new data.
When I exported the model I got the following messages
Variables have been created in the base workspace.
To use the exported classifier 'trainedClassifier' to make predictions on new data, T, use 
  yfit = predict(trainedClassifier, T{:,trainedClassifier.PredictorNames})
If your new data contains any integer variables, then preprocess the data to doubles like this:
  X = table2array(varfun(@double, T(:,trainedClassifier.PredictorNames)));
  yfit = predict(trainedClassifier, X)
So I ran the code below on a vector called toytestdata and got the following
Error using numel
Bad subscripting index.
Error in toyscript (line 9)
yfit = predict(trainedClassifier, T{:,trainedClassifier.PredictorNames})
Here is my code
% code
% Variables have been created in the base workspace.
% To use the exported classifier 'trainedClassifier' to make predictions on new data, T, use 
%   yfit = predict(trainedClassifier, T{:,trainedClassifier.PredictorNames})
% 
% If your new data contains any integer variables, then preprocess the data to doubles like this:
%   X = table2array(varfun(@double, T(:,trainedClassifier.PredictorNames)));
%   yfit = predict(trainedClassifier, X)
    T=toytestdata
    yfit = predict(trainedClassifier, T{:,trainedClassifier.PredictorNames})
      end
My questions are
1.Is my input data in the right format.?
2. Do I need to change the PredictorNames field in my code. I note that this is listed as column_1 in the model…..is there any way to change this..
3. Do I need to do anything else to get this running.
Thanks very much for your help. I attached a few pictures which may assist..
Have a nice weekend
Kevin

Best Answer

To match that code, your T needs to be a table with column named "column_1" (the same as the predictor name)
Related Question