Equation List¶
The Equation List window displays a selection list of the curve-fit equations. This list will initially be sorted by the criteria set in the Curve-Fit Preferences Process menu option. This list can subsequently be sorted by any one of TableCurve 2D's goodness-of-fit sort criteria and its content can be filtered to contain only certain types of equations.
By default, rational equations detected as having a pole in the range of the X-data are initially excluded.
The Equation List window has its own menu.You can also access its content-related functions via the List menu in the Curve-Fit Graph window. The Equation List is toggled on and off by the Eqn. List button in the Curve-Fit Graph window.
The window size and position you choose for the Equation List are automatically saved across sessions. TableCurve 2D saves two Equation List window positions and sizes, one for the Graph Start option and one for the List Start option. Options to save the Equation List to a file, to display the List Window in color, and to print all or part of the list are present only in the List Window's own menu and not in the List menu of the Curve-Fit Graph. To access these functions when the Equation List is too small to have a menu, maximize the Equation List window.
Rank¶
The first item in each line of the equation list will be the ranking of the equation by the goodness-of-fit criteria currently set. Initially, all successfully fitted equations are shown except rational equations which have an undefined region in the X-range of the data. These rationals can be included in the list by toggling the Rationals Fully Defined Xmin-Xmax menu item. The rank will always reflect the position within all fitted equations, regardless of how the list is filtered.
Goodness-of-Fit¶
The second item will be the goodness-of-fit value. This may be the r² coefficient of determination, the DOF-Adjusted r², the Fit Standard Error, the Maximum Error, or the F-statistic. The current goodness-of-fit criteria will be shown in the window's title bar. When the All Goodness-of-Fit in List option in the List menu is checked, all four goodness-of-fit statistics will be displayed within the equation list.
Floating Point Demand¶
The third item will be the FP or floating point demand for the TableCurve 2D generated language code for this equation. A value of 1 corresponds to 1 floating point addition. The higher the number, the greater the execution time for the equation code.
Equation Number and Equation¶
The last items in an equation list entry will be the TableCurve 2D equation number and the equation expression. The format for non-linear equations can be set in the preferences.
Selecting an Equation¶
If you make a selection from the Equation List using a mouse, an equation is selected as soon as you click on it. All Review windows currently open will be updated to reflect the new equation. If the Equation List is the currently active window, you can also move through the list using the cursor keys and make a selection by pressing Enter. If you maximize the Equation List window, regardless of Review startup option, the Equation List window is automatically closed when a selection is made.
File Menu¶
The Save As option is used to save the Equation List to a simple ASCII file, to a Lotus WK1 file, to a Lotus WK3 file, to a fully formatted Rich Text Format (RTF) file, or to a partially formatted Wordstar file which retains superscripts and subscripts. In the case of the ASCII text, WK1, and WK3 files, superscripts are not preserved. A save to the RTF format seeks to exactly preserve the appearance of the equation list when imported into word processing and desktop publishing products. This option preserves columnar formatting, font colors, bolding, superscripts, and subscripts. In order to save a black and white RTF text file, uncheck the Color option in the List menu.
Use the Printer Setup item to select the current printer and optionally to configure the printer's own driver.
The Print item is used to print either the currently displayed contents of the list or the entire list. The Text Print dialog is used to customize the printing of the information.
Edit Menu¶
The Copy option copies the full equation list, memory permitting, to the Windows clipboard as unformatted ASCII text, as formatted Rich Text Format (RTF) text, and in the Lotus WK1 format. The unformatted text format is space delimited and is used when you to paste the equation list into a text-based program. The RTF format seeks to exactly preserve the appearance of the list when pasted into word processing and desktop publishing products. The RTF format preserves columnar formatting, font colors, bolding, superscripts, and subscripts. In order to save a black and white RTF text file, uncheck the Color option in the List menu prior to the Copy. For some products where RTF is the default clipboard format, such as Microsoft Word, use a Paste As, Paste Special, or similar option in order to specifically select the unformatted text. When you paste the equation list into a spreadsheet, the WK1 format is used.
The ASCII Editor option copies as much of the equation list as possible into the TableCurve 2D ASCII editor. On systems with 16-bit edit controls, the maximum size is about 700-800 equations. In all of the Edit menu options, the superscripts in the equation strings will not be preserved.
List Menu¶
If the Keep List After Selection is checked and the Equation List window is not maximized, the Equation list will not be removed after a selection is made. If this Keep List After Selection is not checked, or if the Window is maximized, the Equation List window is closed after each selection.
When the All Goodness-of-Fit in List option is checked, all four goodness-of-fit statistics will be displayed within the equation list. First will be the r² coefficient of determination, then the DOF-Adjusted r², the Fit Standard Error, and finally the F-statistic.
The style items include setting a font for the list. This Font for List font selection is unique to the Equation List. The Color option toggles the display of the list between color and black and white.
Filter Menu¶
All of the Filter Menu items are also in the Curve-Fit Graph's List menu. You may apply eight different filter options to the equation list so that it contains All Equations, only the Simple Equations, only Polynomial/Rational Equations, only Polynomial Equations, only Rational Equations, only Non-Linear Equations, only Continuous Trend First Derivative equations (those which have a continuously increasing or decreasing first derivative across a specified X-range), or Constant Sign First Derivative equations (those which have a constant sign first derivative across a specified X-range). These options reset the selected equation to the first within the list.
For the two derivative scanning options, you can use the Set X Range for Derivative Filtering option to specify the X range as well as how many sampling points are to be used. The default scan will range from Xmin to Xmax+1/2 Xrange with 32 points. The larger the number of sampling points, the likelihood of an equation with an undefined region or pole slipping into the list will be lessened. The procedure is a linear one, meaning that it will take twice as long to scan the list when 64 sampling points are used.
By default, rational functions are not displayed if they have a pole or undefined region in the range of the X data. The Rational Defined Xmin to Xmax option is used to toggle this additional filter.
Sort Menu¶
All of the Sort Menu items are also in the Curve-Fit Graph's List menu. You can sort the current list at anytime by r² coefficient of determination, by DOF-Adjusted r², by Standard Error (Root MSE), by Maximum Error, by F-statistic,.by Floating Point Speed, or by Equation Number. If you want the best possible least squares fit with regard to the number of coefficients, use either the DOF-Adjusted r² or the Standard Error criteria. If you want the equation that represents the best possible model for describing the data, use the F-statistic. The F-statistic sort will often bring many simpler equations nearer the top of the list if the equations do a respectable job describing the data using a modest number of coefficients. A sort resets the selected equation to the first equation in the list.
The Arrange by Parameter Count option uses the coefficient count as the primary sort key, and the selected goodness-of-fit sort criterion as the secondary key. When this option is selected, all of the two parameter fits will be at the top of the list and all of the twenty-one parameter fits will be at the bottom. Within each parameter count grouping, the equations will be sorted by the currently selected goodness-of-fit.