Presentation is loading. Please wait.

Presentation is loading. Please wait.

Sigmoidal Response (knnl558.sas). Programming Example: knnl565.sas Y = completion of a programming task (1 = yes, 0 = no) X 2 = amount of programming.

Similar presentations


Presentation on theme: "Sigmoidal Response (knnl558.sas). Programming Example: knnl565.sas Y = completion of a programming task (1 = yes, 0 = no) X 2 = amount of programming."— Presentation transcript:

1 Sigmoidal Response (knnl558.sas)

2 Programming Example: knnl565.sas Y = completion of a programming task (1 = yes, 0 = no) X 2 = amount of programming experience (months) n = 25

3 Programming Example: input data programming; infile ‘H:\My Documents\Stat 512\CH14TA01.DAT'; input experience complete; proc print data=programming; run; Obsexperiencecomplete 1140 2290 360 4251 5181 640 7 0 8120 9221 1060 11301 12110 ⁞⁞⁞

4 Programming Example: Scatterplot proc sort data=programming; by experience; title1 'scatterplot with smoothing'; symbol1 v=square i=sm60 c=blue; proc gplot data=programming; plot complete*experience; run;

5 Programming Example: Scatterplot (cont)

6 Programming Experience: Logistic Regression proc genmod data=programming descending; model complete = experience/dist=binomial link=logit type1 aggregate; run;

7 Programming Experience: Logistic Regression (cont) PROC GENMOD is modeling the probability that complete='1'. Response Profile Ordered Value complete Total Frequency 1111 2014 Criteria For Assessing Goodness Of Fit CriterionDFValueValue/DF Deviance1719.87941.1694 Scaled Deviance1719.87941.1694 Pearson Chi-Square1719.62261.1543 Scaled Pearson X21719.62261.1543 Log Likelihood-12.7123 Full Log Likelihood-11.3260 AIC (smaller is better)26.6520 AICC (smaller is better)27.1974 BIC (smaller is better)29.0897 Algorithm converged.

8 Programming Experience: Logistic Regression (Cont) Analysis Of Maximum Likelihood Parameter Estimates ParameterDFEstimate Standard Error Wald 95% Confidence Limits Wald Chi- Square Pr > ChiSq Intercept1-3.05971.2594-5.5280-0.59145.900.0151 experience10.16150.06500.03410.28886.180.0129 Scale01.00000.00001.0000 Note:The scale parameter was held fixed.

9 Programming Example: Fitted Curve symbol1 v=square i=none c=blue; symbol2 v=none i=join c=blue; proc gplot data=fit; plot complete*experience pred*experience/overlay; run;

10 Programming Example: Hypothesis test Note:The scale parameter was held fixed. LR Statistics For Type 1 Analysis SourceDevianceDFChi-SquarePr > ChiSq Intercept28.7513 experience19.879418.870.0029


Download ppt "Sigmoidal Response (knnl558.sas). Programming Example: knnl565.sas Y = completion of a programming task (1 = yes, 0 = no) X 2 = amount of programming."

Similar presentations


Ads by Google