tag:blogger.com,1999:blog-3366935554564939610.post6464453933421427662..comments2021-09-20T02:39:30.570-04:00Comments on Data Miners Blog: Multidimensional Chi-Square, Expected Values, Independence, and All That, Part 1Michael J. A. Berryhttp://www.blogger.com/profile/06077102677195066016noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-3366935554564939610.post-89973185289977217912010-04-07T02:38:05.242-04:002010-04-07T02:38:05.242-04:00Hello :)
I am a fresher into data mining
I am into...Hello :)<br />I am a fresher into data mining<br />I am into Discretizing data using chi square test.<br />I have a query here<br />Is there any specific method to divide range of data into intervals???<br />as of now am using heuristic method like >5,<=5,>10,<=10,....... and so on...<br />If some one could help me with this<br />is there any particular standards that the data range should be divided into these many number of intervals ???<br /><br />Thanks and Regards<br />SamaritanAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-3366935554564939610.post-90282324074516850302009-08-25T19:05:47.953-04:002009-08-25T19:05:47.953-04:00Dear Friends,
You wrote the book data analysis wit...Dear Friends,<br />You wrote the book data analysis with SQL... do you recommend the Microsoft NN algorithm? <br />what you thinkl about MSFT NN maturity?<br />regards,<br />Pedro<br />www.pedrocgd.blogspot.comPedroCGDhttp://www.pedrocgd.blogspot.comnoreply@blogger.comtag:blogger.com,1999:blog-3366935554564939610.post-41736180010553160032009-01-23T16:47:00.000-05:002009-01-23T16:47:00.000-05:00Hi,I don't think you intended this blog to be a ch...Hi,<BR/><BR/>I don't think you intended this blog to be a chi-square help page but I have not been able to get an answer to the question below elsewhere so I hope that you can help. Thank you in advance for any assistance you can offer.<BR/><BR/>When conducting a chi-square using dependent observations, what is the appropriate method for calculate the statistic? Example:<BR/><BR/>A randomized-controlled trial (pre- and post- observations in both an intervention and control group). The outcome of interest is "ate peanut butter in the last week" (yes/no).<BR/><BR/>Pre Intervention: N = 250<BR/>(Yes = 70, No = 180)<BR/><BR/>Post Intervention N = 200 (some people lost to follow-up)<BR/>(Yes = 120, No = 80)<BR/><BR/>Pre Control N = 275<BR/>(Yes = 90, No = 185)<BR/><BR/>Post Control N = 230<BR/>(Yes = 100, No = 130)<BR/><BR/>One note: <BR/>(1) there is no access to original data so any cell information has to be back calculated, if possible.<BR/><BR/>(2) I have attempted to calculate using a 2x2x2 chi-square inputing the following data:<BR/>70 180 120 80 90 185 100 130<BR/><BR/>(this represents the cells across each row for the intervention group and then the cells across each row of the control group).<BR/><BR/>I have also tried backing into the cell values using a McNemar's table, but I don't know if that is accurate (and it requires equal sample sizes and doesn't compare between groups, just within). <BR/><BR/>Also, <BR/><BR/>Does the procedure differ depending on whether cohort or serial cross-sectional data are used (i.e., dependent vs. independent observations)?<BR/><BR/>How does one account for differing sample sizes in each group (or at different times of measurement).<BR/><BR/>Again, sorry to bombard you with questions. My internet research has been pretty fruitless thus far.<BR/><BR/>Thank you.fitymihttps://www.blogger.com/profile/01636932128876852238noreply@blogger.com