Using open source node to run simple sklearn model

[DaisyQL]

Hi, 

I have followed the demo and example code to run a simple Random Forest model using Open Source Node in Model Studio.

The code is able to run as Preprocessing node but not Supervised Learning. My gut is telling me I might miss define some required variable in order to render the Assessment, but I can not figure out why, could you help to take a look? Thanks!

from sklearn.ensemble import RandomForestClassifier
import pandas as pd
X = dm_traindf.loc[:, dm_input]
y = dm_traindf[dm_dec_target]

params = {'n_estimators': 100}
dm_model = RandomForestClassifier(**params)
dm_model.fit(X, y)

fullX = dm_inputdf.loc[:, dm_input]
dm_scoreddf = pd.DataFrame(dm_model.predict_proba(fullX), columns=['P__va_d_ped_death_bin0', 'P__va_d_ped_death_bin1'])

More information, the name of the target is "_va_d_ped_death_bin" and it is interval, but it only has 0 and 1 value.  

[DaisyQL · Accepted solution]

Actually, I just figure it out and I was able to run the python model. after this fix:

dm_scoreddf = pd.DataFrame(dm_model.predict(fullX), columns=['P__va_d_ped_death_bin'])

So what happened is I copied the example code without checking. For interval variable "_va_d_ped_death_bin", the script did not recognize the levels. Once I fix how to create the dm_scoreddf it works afterwards