Goodness of Fit for a Logistic Regression
In previous section, we studied about Multiple Logistic Regression
- Classification Matrix
- Accuracy
Classification Table & Accuracy
| Predicted / Actual | 0 | 1 |
|---|---|---|
| 0 | True Positive (TP) | False Positive (FP) |
| 1 | False Negative (FN) | True Negative (TN) |
- Also known as confusion matrix
- \(Accuracy=\frac{(TP+TN)}{(TP+FP+FN+TN)}\)
Classification Table in R
threshold=0.5
predicted_values<-ifelse(predict(prod_sales_Logit_model,type="response")>threshold,1,0)
actual_values<-prod_sales_Logit_model$y
conf_matrix<-table(predicted_values,actual_values)
conf_matrix## actual_values
## predicted_values 0 1
## 0 257 3
## 1 5 202Accuracy in R
accuracy<-(conf_matrix[1,1]+conf_matrix[2,2])/(sum(conf_matrix))
accuracy## [1] 0.9828694
The next post is about multi collinearity an individual impact of variables in logistic regression.
0 responses on "203.2.4 Goodness of fit for Logistic Regression"