Open jun-yan opened 5 years ago
It is unclear what statistical computing method will be used. This seems to be a regression analysis, with variable selection. An EM algorithm works when there are missing data. What's your missing data?
Assume the cost of either smoke and non-smoke are Normal distribution. The missing data is the E(x) and Variance of the distribution. So we will use EM algorithm to calculate both smoke and non-smoke's distribution and compare the difference.
They are your model parameters. When you take conditional expectation in the E-step, what variables are you taking expectation with respect to? Please add more details and notations if necessary. I'd appreciate an explanation.
They are your model parameters. When you take conditional expectation in the E-step, what variables are you taking expectation with respect to? Please add more details and notations if necessary. I'd appreciate an explanation.
We add our ideas in files called"final project, pdf - starting plan".
It is unclear what statistical computing method will be used. This seems to be a regression analysis, with variable selection. An EM algorithm works when there are missing data. What's your missing data?