[FIXED] Predictions after modeling: output array

Issue

After fitting a decision tree I want to do some predictions.

• How can I make this output predicted array more readable ? As you can see, the output corresponds to the features' input values but how can I show the corresponding names above the values? For example the first value is 0.28945 corresponding to a column named time_ms but since i have multiple columns I can't distinguish all of them. How can I display the names above the input values?

• These input values return a classification into [1] corresponding to the value "yes". How can I replace it to display yes instead of 1

• Finally, is it useful (or the best way) to do predictions like this when there are multiple features and not just two just like in the Iris database (petal length, width) ?

clf_dt_grid = grid_search_cv.fit(X_train, y_train)
y_pred_grid = grid_search_cv.predict(X_test)

#Predictions on training set
yhat= clf_dt_grid.predict(X_train)
acc = accuracy_score(y_train,yhat)
print(f'Train predictions accuracy : {acc}%')

#New input for predictions 
new_input = [[0.28945,6593,0,178,2,154.5,True,6593.0,4.0,0,1,2.0,0.00,862.0,0.21,524103.0,2]]
new_output = clf_dt_grid.predict(new_input)
print(new_input,new_output)

>>> [[0.28945, 6593, 0, 178, 2, 154.5, True, 6593.0, 4.0, 0, 1, 2.0, 0.0, 862.0, 0.21, 524103.0, 2]] [1]

Solution

You can put your train/test data into a pandas dataframe and then everything would be more readable.

For example, list out the column names and then use the dataframe constructor like this:

col_names = ['time_ms', 'col2', 'col3',..., 'colN']
train_df = pd.DataFrame(columns=col_names, data=X_train)
test_df = pd.DataFrame(columns=col_names, data=X_test)

Then for your new predictions, you can structure it the same way in a data frame (e.g. call it new_input, then call predict:

new_output = clf_dt_grid.predict(new_input.iloc[0].values) # get the first row of values only

print(new_input.iloc[0], new_output)

Answered By - TC Arlen