Dear all, I'm trying to assess model calibration using modified H–L χ2 statistic according to equation (3.2) or (3.50) written in the following reference: http://www.sciencedirect.com/science/article/pii/S0169716103230017 Now I 'm writing a sas macro to estimate the modified H–L statistic . And I want to check my sas code if there any other sas macros available or published. I would be very grateful if someone would give me information about this. Thanks in advance, Yasu
... View more
Hello,
I'd like to use the text topics object on VA interface. I have the text variable on my dataset, I can select that. I also have an ID column on my dataset, however, I cannot set that as the uniq identifier when trying to apply the text topics object because it says:
No data items in the active data can be set as unique row identifiers.
Here is the relevant SS.
What would be the reason for that and how could I solve this issue?
Regards!
... View more
Hello,
I am building models on VA interface. Is it possible to do a 80/20 train/test split in this interface? Or do I need to use SAS Studio or VDMML pipelines to do that split? Thank you!
... View more
Hi Sas expert team, Please help!! I don't know why in SAS Model manager, the performance report (for logistic regression model) run successfully but cannot show GINI while it still show ROC ( GINI = 2*Area under the curve ROC - 1) In SAS Model manager, I created a project with model "logistic regeresion" by PMML method. This model has target variable "Y" , which is a binary with event value = 1. This project was tested run correctly & successfully. After upload the input data table (sample of 2,746 observations with 18 feature variables and "Y" target variable) for the performance report , I click "run now" and the performance report only show ROC and did not show GINI, which is calculated based on ROC. This is properties of the model/project: These are properties of performance report: This is performance report:
... View more
I have a dataset with five variables I would like to use to display the data on a graph. The data is grouped by the first variable (p) and sub-grouped by the second variable (i) along the x-axis. The fourth variable (nor) determines the height of the point along the y-axis. The fifth variable (mnor) is the mean of the nor values for each subgroup. I want to replace the circle with a unique symbol determined by the third variable (date), so that the graph is easier to interpret the results from different days the data was collected. I have been able to replace the point with the date but would prefer a symbol to minimize clutter. I've included a sample code below. The first graph is my original graph without including the date variable, and the second graph is replacing the points with the date value. data samp;
input p$ i date: date9. nor mnor;
format date date9.;
datalines;
A 0 16May2024 1.230 1.000
A 0 23May2024 0.770 1.000
A 1 16May2024 0.014 0.021
A 1 23May2024 0.028 0.021
B 0 16May2024 1.576 1.504
B 0 23May2024 1.432 1.504
B 1 16May2024 0.064 0.071
B 1 23May2024 0.078 0.071
C 0 16May2024 0.432 0.510
C 0 23May2024 0.588 0.510
C 1 16May2024 0.011 0.009
C 1 23May2024 0.007 0.009
;
title "Example Graph 1";
proc sgplot data = samp noborder;
scatter x = p y = nor /
markerattrs = (symbol=CircleFilled) group = i groupdisplay = cluster;
highlow x = p low = mnor high = mnor / nofill type = bar barwidth = 0.4 group = i groupdisplay = cluster;
xaxis type = discrete labelattrs = (size = 9) display = (nolabel);
yaxis labelattrs = (size = 9) display = (nolabel) grid;
run;
title "Example Graph 2";
proc sgplot data = samp noborder;
scatter x = p y = nor /
markercharattrs = (weight = bold) markerchar = date group = i groupdisplay = cluster;
highlow x = p low = mnor high = mnor / nofill type = bar barwidth = 0.4 group = i groupdisplay = cluster;
xaxis type = discrete labelattrs = (size = 9) display = (nolabel);
yaxis labelattrs = (size = 9) display = (nolabel) grid;
run;
... View more