# binary logistic regression

Six Sigma – iSixSigma › Forums › Old Forums › General › binary logistic regression

- This topic has 4 replies, 5 voices, and was last updated 15 years, 3 months ago by O’Connell.

- AuthorPosts
- January 5, 2004 at 10:26 am #34213

aparna dasParticipant@aparna-das**Include @aparna-das in your post and this person will**

be notified via email.anybody can tell me somthing about “binary logistic regression” and

when it is used for which type of data0January 5, 2004 at 11:06 am #93953

indreshParticipant@indresh**Include @indresh in your post and this person will**

be notified via email.Binary logistic regression is useful in two applications: analyzing the differences among discrete Xs and modeling the relationship between a discrete binary Y and discrete and/or continuous Xs. Binary logistic regression can be used to model the relationship between a discrete binary Y and discrete and/or continuous Xs. The predicted values will be probabilities p(d) of an event such as success or failure-not an event count. The predicted values will be bounded between zero and one (because they are probabilities).

Generally speaking, logistic regression is used when the Ys are discrete and the Xs are continuous

The goodness-of-fit tests, with p-values ranging from 0.312 to 0.724, indicate that there is insufficient evidence for the model not fitting the data adequately. If the p-value is less than your accepted a level, the test would indicate sufficient evidence for a conclusion of an inadequate fit.

rgds,

indresh0January 5, 2004 at 2:34 pm #93956

Bert NijhuisParticipant@Bert-Nijhuis**Include @Bert-Nijhuis in your post and this person will**

be notified via email.Oke, most people are familiar with the least squares method for regression. However, this type of regression i its simple form is only practicle for continuous variables like temperature, pressure, etc.

What if you have discrete (categorical) data like taste, good/false? We have different scales in statistics as binary (yes/no), ordinal (weak, moderate, strong) and nominal (blue, green, black). (Note: The difference between oridnal and nominal is that an ordinal scale has at least some natural order, while nominal has no natural order.)

For characteristics (Y’s) of these scales (binary, ordinal, nominal) we know different types of logistic regression which uses maximum likelihood estimates.0January 6, 2004 at 2:20 pm #93978We have used it to narrow down the factors and model the probabilities of winning/losing proposals based on a variety of factors. The results have been somewhat surprising yet remarkably consistent. A good tool, but it may take a lot of grunt work to get all your data in the right format (it did for us).

0November 18, 2004 at 6:46 pm #110937Breizh, would you mind sharing what you’ve done in more detail? We are embarking on similar analysis and I’d like to see what you’ve done in this area.

~Brian0 - AuthorPosts

The forum ‘General’ is closed to new topics and replies.