ABSTRACT
Logistic regression models are used in applications such as classifying patients who received the flu shot or not; finding factors that differentiate between someone who is obese and someone who is not obese, or predicting whether the patient will be diagnosed with Alzheimer's disease or not based on plaque score. The Bangladesh dataset consists of responses from more than 16,000 women taken from the 2011 Bangladesh Demographic and Health Survey (BDHS). A logistic regression model is a predictive method for binary outcomes modeling. The logit function is useful as a means of transforming the probability to a desirable linear scale. It appears that the probability of using contraceptives and the odds of using contraceptives increase with age. The predicted probabilities in a logistic regression model are obtained as means on a probability scale. A chi-square statistic is used to check the fit of the model for the data in a contingency table.
