Feature Selection using ReliefF function

5 visualizaciones (últimos 30 días)
Nina Perf
Nina Perf el 9 de Sept. de 2021
Comentada: Sulaymon Eshkabilov el 9 de Sept. de 2021
Hello,
I am using the ReliefF feature selection in around 700 features (1 table column is one feature). I am getting an error that I should supply a numeric X. However, the X is always double. The fscmrmr function works well so I don't know why the relieff doesn't.
[idx,weights] = relieff(FeaturesData(:,2:700),ScoreData,10);
% Error using relieff (line 86) You must supply numeric X.
bar(weights(idx));
idx(1:10);
xlabel('Predictor rank');
ylabel('Predictor importance weight');
Informations:
ScoreData: 422×1 double
Values:
Min 1
Median 5
Max 5
Feature1_Mean: 422×1 double
Values:
Min -27.801
Median -0.064483
Max 22.406
Feature2 ...
Thank you in advance!
  1 comentario
Sulaymon Eshkabilov
Sulaymon Eshkabilov el 9 de Sept. de 2021
Could you pl., post your sample (very small portion of it) data table to view what you are inputting for X in relieff() fcn?

Iniciar sesión para comentar.

Respuestas (0)

Categorías

Más información sobre Data Distribution Plots en Help Center y File Exchange.

Productos


Versión

R2021a

Community Treasure Hunt

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

Start Hunting!

Translated by