Logloss in use

Report a typo

Imagine that you want to compare two logistic regression models. Use log-loss as a metric. Below you find the predictions of each model(yiy_i is 0 if the label is 'cat' and 1 if the label is 'dog').

First model:

yiy_i pip_i
1 0.74
0 0.35
0 0.41

Second model:

yiy_i pip_i
1 0.68
0 0.27
0 0.33

Your task is to find the best model and provide the value of its log-loss function. Round the answer to the second decimal place. For example, if your answer is 0.34790.3479, you write 0.350.35.

Tip: LogLoss=1mi=1m(yiln(pi)+(1yi)ln(1pi))\text{LogLoss} = -\frac{1}{m} \cdot \sum\limits_{i=1}^{m} (y_i \cdot \ln(p_i) + (1-y_i) \cdot \ln(1-p_i))

Enter a number
___

Create a free account to access the full topic