home *** CD-ROM | disk | FTP | other *** search
- G R A F P A K
-
- Documentation
-
-
- Grafpak requires a Color/Graphics card and at least 192K of
- RAM to operate. It does not require the use of an 8087 but
- I have a version of Grafpak that does and it is much faster
- than w/out.
-
-
- Grafpak is a X,Y data , analysis and graphing program. The
- user may input pairs of X,Y data, store and retrieve them as
- well as plot them , either in raw form or fitted with 2 types
- of curves. Either Interpolation or a Polynomial is used The
- data may be fitted with an interpolation scheme that makes any
- data smooth. The method is called spliced cubic interpolation
- for those interested. The alternative approach is to fit the
- data with an Nth order Polynomial in the form of Y = NN * X^NN
- + ... N1 * X + N0. This method is a form of smoothing and
- extrapolation good for forecasting and trend analysis. Both
- methods have thier merit.
-
- Grafpak is Menu and Command line driven. The appropriate
- Function key or first letter of a command is all that is
- necessay. Carriage returns are only necessary for numeric
- input and file names.
-
-
- Level #1
-
- The first level is the data input or utility level. This is
- where you tell the program how to get it's data. You may
- either Enter new data , Retrieve a disk file , Read a disk
- directory or Quit.
-
- Entering new X,Y pairs takes you to a new menu where you can
- input new pairs or enter user coefficients. Coefficients are a
- rather advanced subject, and apply to coefficients for a
- polynomial of the Nth order. The same coefficients are
- disaplayed after fitting data with a polynomial. Unless you
- know what you're doing you should let this feature alone.
- Upon entering new pairs you will be taken to the next level
- down where you may operate on your data.
-
- Retrieving a disk file is also straight forward. The file
- must , of course , already exist. The files are created in
- two ways. You may use the "Enter new data option" and enter
- your pairs, after entry you will be prompted to save these
- pairs in a file , you should do so. If you have an editor or
- are patient with the DOS `copy con: filename' you can enter
- data into your own file and then use Retrieve to read your
- data. The file format I use is straight ASCII with the
- first line being the integer number of pairs to follow and
- the next N lines being pairs of data , either integer or FP
- , with a space seperating the pairs (eg 1.0 2.45). After
- sucessfully reading in a file you will be taken to level 2,
- where you can operate on your data.
-
-
-
- Level #2
-
- Level 2 will be your main operating level. At this point
- you have data in memory. Grafpak is based on two arrays,
- where array #1 holds the user's data and array #2 will hold
- some representation of the data in array #1 or a second
- file inported into memory and stored in array #2. So at this
- level you may either `Read a second file into memory' ,
- `Display the data or Edit the data' , `Plot the Data' ,
- `Interpolate the data' , `Polynomial fit to the data' or
- return to level #1.
-
- Reading a second file into memory does as it implies, the
- new data is put into array #2 and the user is returned the
- the level #2 menu again. (See above).
-
- Displaying or Editing the data allows the user to see what
- he has just entered and allows him to Edit it if desirable.
- The user is also given a chance to re-file the edited data
- if edited.
-
- Plotting the data is the whole idea of this program. This
- will call the plotting module. This will line graph the
- user's data in array #1 (also array #2 if a second file has
- been read in). There will be no interpolation or fitting at
- this point just a display of the raw data. Once the plot is
- displayed on the screen the user has a command line of
- choices , Quitting the plotting module will bring the user
- back to the level #2 menu again.
-
-
- Interpolating the data in array #1 will result in an
- interpolation of array #1 in array #2. The user will be
- asked how many intervals he desires between actual data
- points. Generally the larger the number of intervals the
- smoother the data will appear. An interval of 1 does
- absolutely nothing except repeat the data in array #1. Once
- you choose your interpolation the user will see a command
- line of options on the very bottom line of the CRT. this is
- a single key command line that will either file the
- resultant interpolation , interpolate any Y for you , or
- Plot the data vs interpolation.
-
-
- Fitting the data will call a least squares polynomial
- evaluation module, and will place the resultant X,Y pairs into
- array #2. The user will be asked for the Order of the
- polynomial (Up to 15th order). The order of a polynomial means
- the highest power of X that a coefficient is generated for. An
- order of 1 means a straight line is passed through the user's
- data, representing an expression like Y = X^1 * C , where C is
- a unique coiefficient and X is between some predefined range.
- Generally speaking the bumpier your data is the higher the
- order of a polynomial will be required to best fit the data.
- Polynomial evaluation is a science in itself and can be used
- for several applications. Once the user enters the order of
- the fit to be done, the program performs the fit and lists the
- coefficients and thier powers of X as well as a percentage of
- fit. Once the fit is performed the user is provided a command
- line of options, on the last line of the CRT. The options
- available are Expanding the fit to give you more pairs of data
- within the same range of X's using the polynomial expression
- just calculated. Filing the new data in array #2 to disk.
- Reviewing the fit vs the data, where the X , Y , Ycalculated
- , and the Delta are dispalyed for all the user's pairs of
- data. DeltaPlot will plot the difference of the user's data
- and the resultant fit. Plot will plot the user's data vs
- whatever is in array #2 , either expanded fit or fit. The
- user also has the option of quitting to either the top most
- level or the level just above.
-
-
-
-
-
- I realize that this is slim documentation , but like most
- programmers that comes last, and I'm constantly changing
- Grafpak so this will have to suffice for now.
- I would appreciate any comments or suggestions that you have
- regarding Grafpak. I can be reached at...
-
- Peter Worcester
-
- c/o
- Pconn RBBS
- PCSI RBBS
- HACKERS RBBS
-
- Thanks , and I hope you find this program useful.
-
-