A Questionnaire On Logistic Regression

Ananya Jha
May 19, 2019   •  29 views

What is confusion matrix?
How is the output of confusion matrix ?
What is TN,TP,FP,FN? Explain with example.
-------- and ------- occurs when the predicted class matches with the actual class ?
FP and FN occurs when your predicted class ------ with your actual class.
What are Type 1 and Type 2 errors ?
-------- is an intuitive performance measure.
What is accuracy and mis-classification rate ?
Error rate is -------- (mathematical Formulae)
What is precision ?
--------- indicates low false positive rate.
Recall is also called ---------.
What is True Positive Rate?
What is specificity ?
Specificity is also called ------------(TNR)
What is F score ?
-------- takes both False Positive and False negative into account.
When does accuracy works well ?
Give the formula for F1 score.
--------- and ------- tells about ROC.(Area under the curve, confusion matrix)

Accuracy is a good measure if data sets are -------. (symmetric)
F1 is useful if you have ------------ class distribution. (uneven)
------------ and --------- are statistical measure of Binary classification test.
Binary classification test is also called ----------
(classification Function)
Sensitivity is complementary to ------------
Specificity is also called as ---------
There is tradeoff between --------- and ---------.
How to check whether your model is a good model or not?
False Negative is complementary to true positive.
False positive rate is complementary to True Negative.
What can you take as your objective function and what can you do with it ?
Logistic Regression can also be used to predict a ----------- that can assume more than 2 values.(dependent variable)

--------- is the measure of goodness of fit in generalized linear model.(Deviance)
For goodness of fit the pre-requisite is ---------.
In LR-Test we take two model and compare the --------- of one model with the -------- of the other model.(fit)

What effect will it have on likelihood value if we remove the predictor variables from the model and what will happen if we add predictor variables ?

What do you mean by the terms less restrictive and more restrictive ? How does these model fit the data ?

How do we calculate lr test when we have log likelihood of two models ?
Give the expression for lr ?
What does L(m*) denote ?
Here m1 is ------- restrictive model and m2 is less restrictive model.(more, less)

How to calculate degree of freedom of a model?