diff options
Diffstat (limited to 'pkg/utilities/nttools/doc/tlinear.hlp')
-rw-r--r-- | pkg/utilities/nttools/doc/tlinear.hlp | 127 |
1 files changed, 127 insertions, 0 deletions
diff --git a/pkg/utilities/nttools/doc/tlinear.hlp b/pkg/utilities/nttools/doc/tlinear.hlp new file mode 100644 index 00000000..66817fef --- /dev/null +++ b/pkg/utilities/nttools/doc/tlinear.hlp @@ -0,0 +1,127 @@ +.help tlinear Aug2000 tables +.ih +NAME +tlinear -- Fit a linear function to one or two table columns by linear +regression. +.ih +USAGE +tlinear intable outtable xcol ycol +.ih +DESCRIPTION +This task generates fitted Y values and their residuals in two columns. +These columns may be written to an output table, but cannot be written +to STDOUT--only the fit parameters can be written to STDOUT. +If there is more than one table in the input list then a separate fit +is made for each table. + +When a column of weights is used (see 'wcol'), +the weights will be applied when computing the +coefficients of the fit (a, b), +their standard deviations (siga2, sigb2), +and chi squared (chi2), +where the names in parentheses are the headings in +the output printed to STDOUT. +If any row has a weight that is exactly zero, +that row will not be counted in the "pts in fit" value. +The weights will NOT be used when computing +the RMS of the residuals and mean of the residuals +(residual rms, residual mean); +these are unweighted averages +except that rows with exactly zero weight will not be included. +.ih +PARAMETERS +.ls intable [file name template] +A list of input tables containing the columns to be fit. +A fit will be made of the columns specified by the 'xcol' and 'ycol' +parameters. If more than one file name is passed to 'intable', all of +the files must use the same column names. +.le +.ls outtable = STDOUT [file name template] +File names for creating output files, or STDOUT to send output to the screen. +If the value of this parameter is "STDOUT" then the parameters of the fit will +be written to STDOUT preceded by a header line (beginning with #) in tabular +form. +If 'outtable' is not "STDOUT" then the number of file +names must match the number +of names in 'intable', and the fitted Y values and residuals will be written +to an output table with the specified name. The parameters of the fit will +be written to the table header. +.le +.ls xcol [string] +Column name in the input tables to be fit. +The values in this column will be fit for the X axis. +(The same column name is used for each input table.) If a name is not specified +for the X values then row number is used. The values in the 'xcol' column will +be copied to 'outtable' unless the output is being directed to STDOUT. +.le +.ls ycol [string] +Column name in the input tables containing value to be fit for the Y axis. +(The same column name is used for each input table.) Values in 'ycol' will +be copied to 'outtable' unless 'outtable = STDOUT'. +.le +.ls (wcol) [string] +Column name in 'intable' that contains weight values for X and Y. +(The same column name is used for each input table.) If no column +name is passed to either the 'wcol' or 'scol' parameters, then a weight +of 1. is used. The value of the 'wcol' column is copied to 'outtable' unless +'outtable = STDOUT'. +.le +.ls (scol) [string] +Column in 'intable' containing the standard deviation of X and Y. +The X and Y values are weighted by the values in 'scol' +as the reciprocal of the values squared. (The same column name is used for each +input table.) If no value is passed to 'wcol' or 'scol', then +a weight of 1. is used. This task can accept either a weight value or a +standard deviation value, but not both. If both 'wcol' and 'scol' are +specified, then the weight column (i.e., 'wcol') will be used. +The value in the 'scol' column is written to 'outtable' unless 'outtable' += STDOUT. +.le +.ls (rows = "-") [string] +Range of rows to use for fitting the data. +The default "-" means that all rows are used. +(Type "help xtools.ranges" for more information.) +.le +.ls (outcoly = "yfit") [string] +Column name for fitted Y values. +This parameter is not used if 'outtable' = STDOUT. +This column will be double data type. +.le +.ls (outcolr = "yres") [string] +Name of the column to contain residuals. +This parameter is ignored if 'outtable' = STDOUT. +This column will be of double data type. +.le +.ih +EXAMPLES +1. Fit the values in the "flux" column in every table whose name begins with +"hr"; put all parameters of the fits in the ASCII file "fit.lis". + +.nf + tt> tlinear hr*.tab STDOUT "" flux > fit.lis +.fi + +2. Generate the same fits as in the previous example, but put the +results in tables, one output for each input table. For example, +the fitted Y values and +residuals for an input table named "hr465.tab" would be put in "hr465h.tab". + +.nf + tt> tlinear hr*.tab hr*%%h%.tab "" flux +.fi + +3. Fit the values in the "flux" column as a function of the values in the +"wavelength" column and write all the parameters of the fit to STDOUT. + +.nf + tt> tlinear hr*.tab STDOUT wavelength flux +.fi +.ih +BUGS +.ih +REFERENCES +This task was written by Betty Stobie. +.ih +SEE ALSO +ranges +.endhelp |