lightgbm how to predict_proba?

Question

if i use lightgbm

there is two methods of using lightgbm. first method: -

model=lgb.LGBMClassifier()
model.fit(X,y)
model.predict_proba(values)

i can get predict_proba method to predict probabilities. if i use it natively

import lightgbm as lgb
dset = lgb.Dataset(X, label=y)
params = {}
params['learning_rate'] = 0.003
params['boosting_type'] = 'gbdt'
params['objective'] = 'binary'
params['metric'] = 'binary_logloss'
model = lgb.train(params, dset, 100)

I can get predict method but predict_proba method dont exist if i use this method. anyone can advise what i did wrong?

Please create a [minimal, reproducible example](https://stackoverflow.com/help/minimal-reproducible-example). That would mean 1. mentioning the version of `lightgbm` you are using 2. including the specific text of any logs or errors that have led you to conclude "predict_proba method doesn't exist" 3. modifying the provided code so anyone can run it (e.g., including code to generate `X` and `y`). — James Lamb, Nov 28 '21 at 23:55

ferdy · Answer 1 · 2021-12-17T06:20:03.027

0

You can only use predict_proba in sklearn fit(). It does not exist in lightgbm native train().

edited Dec 17 '21 at 06:20

answered Dec 15 '21 at 19:40

ferdy

4,396
2
4
16

whats the workaround for using native api to predict probabilities then? – desmond Dec 16 '21 at 05:09
At the moment the solution is to use fit(). – ferdy Dec 16 '21 at 08:25
fit just, well.. fit the data, after that would be to predict isnt it? – desmond Dec 17 '21 at 05:58
When you use the sklearn fit(), you will have access to predict() and predict_proba(). But you have to prepare your datasets and parameters that are acceptable to sklearn. – ferdy Dec 17 '21 at 06:10

lightgbm how to predict_proba?

1 Answers1