HISTOGRAM Statement: CAPABILITY Procedure

Output Data Sets

You can create two output data sets with the HISTOGRAM statement: the OUTFIT= data set and the OUTHISTOGRAM= data set. These data sets are described in the following sections.

OUTFIT= Data Set

The OUTFIT= data set contains the parameters of fitted density curves, information about chi-square and EDF goodness-of-fit tests, specification limit information, and capability indices based on the fitted distribution. Since you can specify multiple HISTOGRAM statements with the CAPABILITY procedure, you can create several OUTFIT= data sets. For each variable plotted with the HISTOGRAM statement, the OUTFIT= data set contains one observation for each fitted distribution requested in the HISTOGRAM statement. If you use a BY statement, the OUTFIT= data set contains several observations for each BY group (one observation for each variable and fitted density combination). ID variables are not saved in the OUTFIT= data set.

The OUTFIT= data set contains the variables listed in Table 5.24. By default, an OUTFIT= data set contains _MIDPT1_ and _MIDPTN_ variables, whose values identify histogram intervals by their midpoints. When the ENDPOINTS= or NENDPOINTS option is specified, intervals are identified by endpoint values instead. If the RTINCLUDE option is specified, the variables _MAXPT1_ and _MAXPTN_ contain upper endpoint values. Otherwise, the variables _MINPT1_ and _MINPTN_ contain lower endpoint values.

Table 5.24: Variables in the OUTFIT= Data Set

Variable

Description

_ADASQ_

Anderson-Darling EDF goodness-of-fit statistic

_ADP_

p-value for Anderson-Darling EDF goodness-of-fit test

_CHISQ_

chi-square goodness-of-fit statistic

_CP_

generalized capability index $C_ p$ based on the fitted curve

_CPK_

generalized capability index $C_{pk}$ based on the fitted curve

_CPL_

generalized capability index CPL based on the fitted curve

_CPM_

generalized capability index $C_{pm}$ based on the fitted curve

_CPU_

generalized capability index CPU based on the fitted curve

_CURVE_

name of fitted distribution (abbreviated to 8 characters)

_CVMWSQ_

Cramer-von Mises EDF goodness-of-fit statistic

_CVMP_

p-value for Cramer-von Mises EDF goodness-of-fit test

_DF_

degrees of freedom for chi-square goodness-of-fit test

_ESTGTR_

estimated percent of population greater than upper specification limit

_ESTLSS_

estimated percent of population less than lower specification limit

_ESTSTD_

estimated standard deviation

_EXPECT_

estimated mean

_K_

generalized capability index K based on the fitted curve

_KSD_

Kolmogorov-Smirnov EDF goodness-of-fit statistic

_KSP_

p-value for Kolmogorov-Smirnov EDF goodness-of-fit test

_LOCATN_

location parameter for fitted distribution. For the Gumbel, inverse Gaussian, and normal distributions, this is either the value of $\mu $ specified with the MU= option or the value estimated by the procedure. For all other distributions, this is either the value specified or estimated according to the THETA= option, or zero.

_LSL_

lower specification limit

_MAXPT1_

upper endpoint of first interval used to calculate the value of the chi-square statistic.

_MAXPTN_

upper endpoint of last interval used to calculate the value of the chi-square statistic.

_MIDPT1_

midpoint of first interval used to calculate the value of the chi-square statistic. This is the leftmost interval that contains at least one value of the variable.

_MIDPTN_

midpoint of last interval used to calculate the value of the chi-square statistic. This is the rightmost interval that contains at least one value of the variable.

_MINPT1_

lower endpoint of first interval used to calculate the value of the chi-square statistic.

_MINPTN_

lower endpoint of last interval used to calculate the value of the chi-square statistic.

_OBSGTR_

observed percent of data greater than upper specification limit

_OBSLSS_

observed percent of data less than the lower specification limit

_PCHISQ_

p-value for chi-square goodness-of-fit test

_SCALE_

value of scale parameter for fitted distribution. For the lognormal distribution, this is the value of $\zeta $ specified or estimated according to the ZETA= option. For all other distributions, this is the value specified or estimated according to the SIGMA= option.

_SHAPE1_

value of shape parameter for fitted distribution. For the beta, gamma, generalized Pareto, and power function distributions, this is the value of $\alpha $, either specified with the ALPHA= option or estimated by the procedure. For the lognormal distribution, this is the value of $\sigma $, either specified with the SIGMA= option or estimated by the procedure. For the Weibull distribution, this is the value of c, either specified with the C= option or estimated by the procedure. For the Johnson $S_ B$ and $S_ U$ distributions, this is the value of $\delta $, either specified with the DELTA= option or estimated by the procedure. For distributions without a shape parameter (Gumbel, normal, exponential, and Rayleigh distributions), _SHAPE1_ is set to missing.

_SHAPE2_

value of shape parameter for fitted distribution. For the beta distribution, this is the value of $\beta $, either specified with the BETA= option or estimated by the procedure. For the Johnson $S_ B$ and $S_ U$ distributions, this is the value of $\gamma $, either specified with the GAMMA= option or estimated by the procedure. For all other distributions, _SHAPE2_ is set to missing.

_TARGET_

target value

_USL_

upper specification limit

_VAR_

variable name

_WIDTH_

width of histogram interval


OUTHISTOGRAM= Data Set

The OUTHISTOGRAM= data set contains information about histogram intervals. Since you can specify multiple HISTOGRAM statements with the CAPABILITY procedure, you can create multiple OUTHISTOGRAM= data sets.

The data set contains a group of observations for each variable plotted with the HISTOGRAM statement. The group contains an observation for each interval of the histogram, beginning with the leftmost interval that contains a value of the variable and ending with the rightmost interval that contains a value of the variable. These intervals will not necessarily coincide with the intervals displayed in the histogram since the histogram may be padded with empty intervals at either end. If you superimpose one or more fitted curves on the histogram, the OUTHISTOGRAM= data set contains multiple groups of observations for each variable (one group for each curve). If you use a BY statement, the OUTHISTOGRAM= data set contains groups of observations for each BY group. ID variables are not saved in the OUTHISTOGRAM= data set.

The OUTHISTOGRAM= data set contains the variables listed in Table 5.25. By default, an OUTHISTOGRAM= data set contains the _MIDPT_ variable, whose values identify histogram intervals by their midpoints. When the ENDPOINTS= or NENDPOINTS option is specified, intervals are identified by endpoint values instead. If the RTINCLUDE option is specified, the _MAXPT_ variable contains an interval’s upper endpoint value. Otherwise, the _MINPT_ variable contains the interval’s lower endpoint value.

Table 5.25: Variables in the OUTHISTOGRAM= Data Set

Variable

Description

_COUNT_

number of variable values in histogram interval

_CURVE_

name of fitted distribution (if requested in HISTOGRAM statement)

_EXPPCT_

estimated percent of population in histogram interval determined from optional fitted distribution

_MAXPT_

upper endpoint of histogram interval

_MIDPT_

midpoint of histogram interval

_MINPT_

lower endpoint of histogram interval

_OBSPCT_

percent of variable values in histogram interval

_VAR_

variable name


OUTKERNEL= Output Data Set

An OUTKERNEL= data set contains information about kernel density estimates requested with the KERNEL option. Because you can specify multiple HISTOGRAM statements with the CAPABILITY procedure, you can create multiple OUTKERNEL= data sets.

An OUTKERNEL= data set contains a group of observations for each kernel density estimate requested with the HISTOGRAM statement. These observations span a range of analysis variable values recorded in the _VALUE_ variable. The procedure determines the increment between values, and therefore the number of observations in the group. The variable _DENSITY_ contains the kernel density calculated for the corresponding analysis variable value.

When a density curve is overlaid on a histogram, the curve is scaled so that the area under the curve equals the total area of the histogram bars. The scaled density values are saved in the variable _COUNT_, _PERCENT_, or _PROPORTION_, depending on the histogram’s vertical axis scale, determined by the VSCALE= option. Only one of these variables appears in a given OUTKERNEL= data set.

Table 5.26 lists the variables in an OUTKERNEL= data set.

Table 5.26: Variables in the OUTKERNEL= Data Set

Variable

Description

_C_

standardized bandwidth parameter

_COUNT_

kernel density scaled for VSCALE=COUNT

_DENSITY_

kernel density

_PERCENT_

kernel density scaled for VSCALE=PERCENT (default)

_PROPORTION_

kernel density scaled for VSCALE=PROPORTION

_TYPE_

kernel function

_VALUE_

variable value at which kernel function is calculated

_VAR_

variable name