Logistic Regression_Optional Lab_Sigmoid Fucntion

Can anyone please explain, how it is still predicting right classes? My intuition says it is not,

Given the graph, out of the 9 samples, it predicts 1 True Positive, 1 False Positive, 4 True Negative, and 3 False Negative. Therefore, it predicts 5 (out of 9) samples correctly.


I think the error rate is high here because you have an outlier here (Encircled in red).

No model would be able to predict this one correctly because it doesn’t correlate with the rest of the data.