There are two main routines exported from the DLL, one of which calculates the (general) one-sided confidence limits, and the other calculates the two-sided limits. You should be able to call these functions from just about anything in Windows 95/98/NT (an example using Visual Basic in Microsoft Excel is included below). The tables below summarize the argument lists; they are long since they are both for the most general case.

Subroutine TwoTail (m, s, nu, alpha, lambda, gammasquared, LowerLimit, LowerTheH, UpperLimit, UpperTheH, TheTime, LowerGuess1, LowerGuess2, UpperGuess1, UpperGuess2, ItNum1, ItNum2, ArgTop, Tolerance, ResultTol, MaxNumberIterations)
Inputs	Type	Comment	Range/Suggested Value for Inputs	If Calculating Confidence Limits for the Mean of a Lognormal Distribution, then you should use...
m	Double	Sample mean	None	Sample mean of natural logarithm of values
s	Double	Sample standard deviation	0.1 to 10.0	Sample standard deviation of natural logarithm of values
nu	Double	We are assuming that s2 is distributed as s2/n times chi-square with n degrees of freedom	2 to 1000	Number of samples minus one
alpha	Double	Desired level of confidence	Between 0 and 1
lambda	Double	Calculated confidence limits are for m + l s2	None	0.5
gammasquared	Double	We are assuming that m is distributed as a normal distribution with mean m and standard deviation s2/g2	nu+1	Number of samples
ArgTop	Double	Upper bound for Bessel function evaluation	800	800
Tolerance	Double	Absolute tolerance for calculated limits	1e-8	1e-8
ResultTol	Double	Additional tolerance term used for concluding convergence if |f(x)| is small enough in some cases; if less than or equal to 0 then this is not used.	-1	-1
MaxNumberIterations	Long	Maximum number of iterations	200	200
Outputs
returnLowerLimit	Double	Calculated lower two-sided limit of level alpha		Lower limit for the mean is exp(returnLowerLimit)
returnLowerTheH	Double	Standardized lower two-sided limit of level alpha
returnUpperLimit	Double	Calculated upper two-sided limit of level alpha		Upper limit for the mean is exp(returnLowerLimit)
returnUpperTheH	Double	Standardized upper two-sided limit of level alpha
returnTheTime	Double	Time required to calculate limits (seconds)
returnItNum1	Long	Number of iterations required to calculate lower limit
returnItNum2	Long	Number of iterations required to calculate upper limit

Example Calling These Functions Using Visual Basic in Microsoft Excel

In a Visual Basic module in Excel, add the code shown below the table. It declares the DLLs (you'll need to probably change the path, depending on where you put the DLL), and defines wrapper routines that pass some of the default options through to the subroutines in the DLL. In a worksheet in the workbook containing the module, you can call these functions by using formulas of the form

where the arguments can be constants (e.g., "=upperconflimit(0.0, 1.0, 10,0.95)") or references to other cells.

Option Explicit


' Declaring the DLLs; replace  the path with whereever you put the dll.  If you put the DLL in
'   the windows\system directory, then you can use just the filename without the path.



Public Declare Sub TwoTail Lib "c:\ConfLimit\Program\Conf95DLL\Conf95DLL.dll" ( _
                mu As Double, _
                sigma As Double, _
                nu As Double, _
                alpha As Double, _
                Lambda As Double, _
                GammaSq As Double, _
                argArgTop As Double, _
                argTolerance As Double, _
                argResultTol As Double, _
                argNumIt As Long, _
                returnLowerLimit As Double, _
                returnLowerTheH As Double, _
                returnUpperLimit As Double, _
                returnUpperTheH As Double, _
                returnTheTime As Double, _
                returnItNum1 As Long, _
                returnItNum2 As Long)


'                returnLowerGuess1 As Double, _
                returnLowerGuess2 As Double, _
                returnUpperGuess1 As Double, _
                returnUpperGuess2 As Double, _


Public Declare Sub OneTail Lib "c:\ConfLimit\Program\Conf95DLL\Conf95DLL.dll" ( _
                WhichLimit As Long, _
                mu As Double, _
                sigma As Double, _
                nu As Double, _
                alpha As Double, _
                Lambda As Double, _
                GammaSq As Double, _
                BesselFunctionArgTop As Double, _
                Tolerance As Double, _
                Tol As Double, _
                MaximumIterations As Long, _
                returnTheH As Double, _
                returnConfLimit As Double, _
                returnTheTime As Double, _
                returnNumberofIterations As Long)

' A wrapper function for calling the routine that calculates the two-sided confidence limits for the case lambda=0.5; this returns the upper limit
Function TwoSidedConfLimit(m As Double, s As Double, nu As Double, alpha As Double, WhichLimit As Long) As Double

' m     : sample mean
' s     : sample standard deviation (s>0)
' nu    : degrees of freedom (nu>0)
' alpha : desired confidence level (01  : return lower limit
    
    Dim mu As Double, _
        sigma As Double, _
        Lambda As Double, _
        GammaSq As Double, _
        returnLowerLimit As Double, _
        returnLowerTheH As Double, _
        returnUpperLimit As Double, _
        returnUpperTheH As Double, _
        returnTheTime As Double, _
        returnItNum1 As Long, _
        returnItNum2 As Long


'        returnLowerGuess1 As Double, _
        returnLowerGuess2 As Double, _
        returnUpperGuess1 As Double, _
        returnUpperGuess2 As Double, _


Dim BesselFunctionArgTop As Double
Dim Tolerance As Double
Dim ResultTol As Double
Dim MaximumIterations As Long

BesselFunctionArgTop = 800#
Tolerance = 0.0001
ResultTol = -1#
MaximumIterations = 200

        mu = m
        sigma = s
        Lambda = 0.5
        GammaSq = (nu + 1#)

     

    Call TwoTail(mu, _
                sigma, _
                nu, _
                alpha, _
                Lambda, _
                GammaSq, _
                BesselFunctionArgTop, Tolerance, ResultTol, MaximumIterations, _
                returnLowerLimit, _
                returnLowerTheH, _
                returnUpperLimit, _
                returnUpperTheH, _
                returnTheTime, _
                returnItNum1, _
                returnItNum2)
                
'returnLowerGuess1, _
                returnLowerGuess2, _
                returnUpperGuess1, _
                returnUpperGuess2, _

                
If WhichLimit = 1 Then
    TwoSidedConfLimit = returnUpperLimit
Else
    TwoSidedConfLimit = returnLowerLimit
End If

End Function 'Function TwoSidedConfLimit



' A wrapper function for calling the routine that calculates one-sided confidence limits in the
' special case lambda=0.5, gamma^2=nu+1
'
' If m and s are the sample mean and standard deviation of the natural logarithm of samples from a lognormal distribution, calculated from
'  (nu+1) samples, then exp(UpperConfLimit) is an upper confidence limit of level alpha for the mean of the lognormal distrubution
'
Function UpperConfLimit(m As Double, s As Double, nu As Double, alpha As Double) As Double

' m     : sample mean
' s     : sample standard deviation (s>0)
' nu    : degrees of freedom (nu>0)
' alpha : desired confidence level (0
       
      
 
       
      

      

  
 
    
 Home 
    
    
 Download 
    
    
 Run 
      Online 
      

        
 
          
 General     
          
          
 Lognormal 
            Mean 
        
      
    
    
 Program 
      Verification