CODEWITHSUNDEEP
pythonpandasmachine-learningxgboost
kalle
kalle

Reputation: 435

Using XGBoost to predict importance or percentage based on inputs

I am working on interpreting my XGBoost model. Take for example, the two datasets trainInput and trainOutput below, respectively:

df.trainInputs
    input1 input2 input3
0      1     0      0
1      1     1      0
2      0     1      1
..

df.trainOutputs
      output
0       1   
1       0
2       1
...

The user inputs have been one-hot encoded and the output data is a list of user output patterns. I am training my XGBoost model with these and then predicting based on another matrix of one-hot encoded user input data from another dataset. I am hoping to retrieve a percentage score for each element of the column, but upon running my model, am only receiving binary output. Is there something that I am missing in building my model? The relevant code:

df.predictInputs
    input1 input2 input3
0      1     1      0
1      1     0      0
2      1     0      1
..

model = xgb.XGBClassifier()
model.fit(trainInputs, trainOutput)
y_pred = model.predict(predictOutput)

Upvotes: 2

Views: 1797

Answers (1)

YLJ
YLJ

Reputation: 2996

If you want to get the result probability (percentage score for each element), use predict_proba instead of predict.

Upvotes: 4

Related Questions