NAME

twodpolyfit -- Fit a polynomial function of temperature and epoch.

USAGE

twodpolyfit intable outtable scheme threshold order1 order2 yname yerrname x1name x2name base_temp base_time nattrib attrib1 attrib2 attrib3 attrib4 attrib5

DESCRIPTION

This task will fit any quantity (for example, dark count), as polynomial functions of temperature and epoch. This routine is usually used to obtain the 16 polynomial coefficients that the Routine Science Data Processing (RSDP) system needs to calibrate the dark signal, pre-amp noise, current to voltage converter (CVC) offset, and other quantities. This task fits the 2-dimensional polynomial function in both dimensions at the same time by using the singular value decomposition (SVD) method. Although the current pipeline only requires third order polynomials in each dimension, this task is a more generalized and flexible tool which can fit polynomials up to 6 orders in each dimension, allowing the user-definable column names in the input table to be used for other purposes.

Input data are single precision, but double precision calculations are performed. Epoch is in double precision throughout. Any number of input data points can be used.

PARAMETERS

intable [file name]

Name of the input table; this table has the following columns:

(yname)		Measured quantity to be fitted with a polynomial 
		function, e.g., dark counts (real).  The actual
		column name is passed from the 'yname'
		parameter.
(yerrname)      Standard deviations of the measured quantity 
		(real).  The actual column name is passed from
		the 'yerrname' parameter.  
(x1name)        Independent variable of the first dimension, 
		usually the temperature (real).  The actual
		column name is passed from the 'x1name' parameter.
(x2name)        Independent variable of the second dimension, 
		usually the epoch (double).  The actual column
		name is passed from the 'x2name' parameter.
'TYPE'          Data type of the observation (char*7).
(attribi)       Attributes to be copied directly from the
		input table to the output table.  The actual 
		column names and data types are passed from the
		parameters 'attrib1' through 'attrib5'.

outtable [file name]

Name of the output table created by twodpolyfit; this table has the following columns:

'BASE_VALUE'    Base value of the calibration coefficients (real).
'A00', 'A01', etc.
                Calibration polynomial coefficients (real).
'SIGMA00', 'SIGMA01', etc.
		Uncertainties of 'A00', 'A01', etc. (real).
'C(10,00)', 'C(20,00)', etc.
		Covariances between 'A10', 'A00', and 'A20', 'A00', 
		etc. (real).
'CHISQ'         Chi-square of the polynomial fit (real).
'PROBABILITY'	The chi-square probability of the fit (real).
'BASE_TEMP'     Reference temperature (real).
'TEMPMIN'       Lower limit of temperature (real).
'TEMPMAX'       Upper limit of temperature (real).
'BASE_TIME'     Reference epoch (double).
'EPOCHMIN'      Lower limit of epoch (double).
'EPOCHMAX'      Upper limit of epoch (double).
'ORDER1'        Order of the polynomial fit in the first 
		dimension (int).
'ORDER2'        Order of the polynomial fit in the second 
		dimension (int).
'NPOINTS'       Number of valid input data points (int).
'TYPE'          Data type of the observation (ch*7).
(attribi)       Attributes to be copied from the input table to
                the output table as passed from the 'attrib1'
		through 'attrib5' parameters.

scheme [string, allowed values: auto | exam | fixed]

Fit scheme to be used in this task. If this parameter is passed a value of "fixed", the fit will be carried out once at the specified polynomial orders and the results will be written to the output table. If the value "exam" is passed, then the fit will be carried out to all orders no larger than the specified orders, a matrix of chi-square probabilities and a matrix of chi-squares are output to the screen and the log file, and no result is written to the output table. If the value "auto" is passed to scheme, all lower orders are carried out in the same way as the "exam" mode, and the lowest order that has a chi-square probability larger than threshold will be used for the final result. If no order has a probability larger than threshold, the specified orders are used.

threshold = 0.[real, min = 0, max = 1]

The minimum chi-square probability used in "auto" mode. (See the description of the scheme parameter for more information).

order= 3 [integer, min = 0, max = 6]

The maximum degree of the fitting polynomial in the first dimension. If this task is being used for one-dimensional fitting, set order2 = 0.

order= 3 [integer, min = 0, max = 6]

The maximum degree of the fitting polynomial in the second dimension. If using this task for one-dimensional fitting, set order2 = 0.

yname = "DOBJ" [string]

Column name of the quantity to be fitted.

yerrname = "DOBJ_ERR" [string]

Column name of the standard deviation of the quantity to be fitted.

x1name = "DET_TEMP" [string]

Column name of the independent variable in the first dimension.

x2name = "EPOCH" [string]

Column name of the independent variable in the second dimension.

base_temp = 0. [real]

Reference temperature in degrees Celsius.

base_time = 32000. [double]

Reference epoch in Julian days since 1900.0.

nattrib = [integer, min = 0, max = 5]

Number of attributes to be copied from the input table to the output table.

attrib1, attrib2, attrib3, attrib4, attrib[strings]

Attributes to be copied directly from the input table to the output table. The following attributes are allowed to be passed to these parameters:

DET_NUM     Detector ID (int).
APER_NAME   Aperture name (char*10).
VOLTAGE     High voltage setting (real).
VGAIND      Gain setting (real).
THRESH      Discriminator threshold setting (real).

EXAMPLES

1. Fit third order polynomials of temperature and time to the quantity "DOBJ" of the input table xtwodpolyfit$input1 and write the results in the table ytwodpolyfit$output1. The standard deviation of "DOBJ" is in the column "DOBJ_ERR", temperature is in "DET_TEMP", and time is in "EPOCH". The reference temperature is 0 and the reference epoch is 0. Copy detector ID's and aperture names from the input table to the output table.

hs> twodpolyfit intable="xtwodpolyfit$input1"
       outtable="ytwodpolyfit$output1" scheme="fixed"
       order1=3 order2=3 yname="DOBJ" yerrname="DOBJ_ERR"
       x1name="DET_TEMP" x2name="EPOCH" base_temp=0 
       base_time=0 nattrib=2 attrib1="DET_NUM" 
       attrib2="APER_NAME"

BUGS

REFERENCES

SEE ALSO