A practical and example filled tour of simple and multiple regression techniques (linear, logistic, and Cox PH) for estimation, adjustment and prediction.

Loading...

来自 约翰霍普金斯大学 的课程

Statistical Reasoning for Public Health 2: Regression Methods

72 评分

A practical and example filled tour of simple and multiple regression techniques (linear, logistic, and Cox PH) for estimation, adjustment and prediction.

从本节课中

Introduction and Module 1A: Simple Regression Methods

In this module, a unified structure for simple regression models will be presented, followed by detailed treatises and examples of both simple linear and logistic models.

- John McGready, PhD, MSAssociate Scientist, Biostatistics

Bloomberg School of Public Health

So in this section, we're going to take on

simple logistic regression which is very similar in spirit to

simple linear regression and

the general linear model framework we set up at the beginning of Lecture One.

In this situation, however,

we're dealing with a binary outcome.

And we're going to end up for reasons that I'll

get into in the second part of this lecture

set that we're going to have to end up modeling it on the log odd scale.

So what we'll be doing is starting with something that's measured as a yes or no,

one or zero, and then transforming it to

a probability then an odds then the log of an odds.

We think we're going to estimate is a linear function of

our predictors is the long odds of the outcome.

And that may seem strange at first but it's not

a particularly convenient scale but we'll see that the result the estimates

we get are one step removed from scales that we're familiar with in

terms of comparisons and making statements about risk through the odds.

And we'll see, we're also not constrained to odds and odds ratios which are

the immediate results we get from logistic regression but with a little work,

we can translate our estimates into proportions or probabilities as well.

In general, though, the generalized framework will be very similar

to what we did with linear regression in terms of the comparisons

being made by our slopes

and the whole spirit of modeling or outcomes and the linear function of predictors.