How do I fit to a step function?

Hi,
I'm using the Levenberg-Marquard algorithm to optimize a fit to some data.  I would like to fit to a step function, but when I try to do this with simulated data, the algorithm is not at all responsive to optimizing the position of where the step occurs.  (It simply returns the initial guess position of the step location, rather than optimizing it.)  I have tried using both the step(x) and sign(x) functions but both have the same problem.  Does anyone have any experience with this? 
Cheers,
Nancy

The problem is with the partial derivatives for the position parameter. If the position changes less than the point spacing, the partial derivative with respect to that parameter is always zero.
I have quickly hacked the ABX subVI to use something larger than the x spacing of the points as increment for the partial derivative (instead of 1e-6), and it converges just fine after that if the initial estimate is reasonable. (See image). Still, I recommend to not do that!
Follow Lynn's advice.
Message Edited by altenbach on 05-10-2006 01:01 PM
LabVIEW Champion . Do more with less code and in less time .
Attachments:
LevMarStep.png ‏205 KB

Similar Messages

  • How can I create a custom step that does nothing but make calls to a dll

    What I'd like to end up with is a Custom step that
    1. Takes several string parameters
    2. Makes several calls to functions in a DLL
    3. Returns and stores error codes returned from the function
    I've been through the KB article "How Do I Make a Custom Step Type".
    This got me about halfway there; while helpful, it didn't provide a
    reference for some of the questions I had, specifically:
    1. What sort of sequence action type should I start with (to use as a template)?
    2. If I start with a "Call Executable", do I need to specify the module when I incorporate the step into the sequence?
    3. How does one pass parameters in and out of a custom step type?
    4. How does the step type report progress/errors to the report?
    Thanks in advance.
    Mike

    Hi,
    To interface to a DLL you will need to use either the DLL Adapter or the CVI Adapter, not the Call Executable adapter.
    Have you been through the Using LabWindowsCVI with TestStand manual. This should answer all your questions, all the chapters are relevant.
    Regards
    Ray Farmer
    Regards
    Ray Farmer

  • How to insert a code for a function module into a Customer Exit Variable?

    I have two Key Figures viz., Net Prchs Rtl, and Net Prchs Unt. Both these Key figures have This Week (TW) and Last Week (LW). There is a variable for This week but there is no variable defined for Last week.
    I need to get data in the column LW (Last Week) for both the key figures.
    In function module EXIT_SAPLRRS0_001 one of the functions I have is:
    Get the previous Fiscal Week
              CALL FUNCTION 'DATE_TO_PERIOD_CONVERT'
    I need to create a new variable to get values in the Last week column for different key figures and use function “'DATE_TO_PERIOD_CONVERT'” in that variable.
    Can anyone please explain me the steps as to how to use a function module in a variable so that when the variable is used in a key figure it shows the output.
    In other words what I want to know is after creating a Customer exit variable of type Characteristic value how do I refer the above mentioned function moduel and insert the code for the function module into the Customer exit variable that I created.
    Thank you.
    TR.

    Hi Wond,
    Thanks a lot for your reply. I understand what you mean but I have never done this before so can you please explain it in a detailed manner. I have the following code:
    Get the previous Fiscal Week
              CALL FUNCTION 'DATE_TO_PERIOD_CONVERT'
                EXPORTING
                  i_date         = ld_date
                  i_periv        = lc_periv
                IMPORTING
                  e_buper        = ln_poper
                  e_gjahr        = ln_bdatj
                EXCEPTIONS
                  input_false    = 1
                  t009_notfound  = 2
                  t009b_notfound = 3
                  OTHERS         = 4.
              IF sy-subrc <> 0.
                MESSAGE ID sy-msgid TYPE sy-msgty NUMBER sy-msgno
                WITH sy-msgv1 sy-msgv2 sy-msgv3 sy-msgv4.
              ELSE.
                lc_poper = ln_poper.
                CONCATENATE ln_bdatj lc_poper+1(2) INTO wa_e_t_range-low.
                wa_e_t_range-sign = 'I'.
                wa_e_t_range-opt = 'EQ'.
                APPEND wa_e_t_range TO e_t_range.
              ENDIF.
            ENDIF.
          ENDIF.
        ENDIF.
    I believe I should now use the above code in my customer exit variable. If that's right can you please explain me the steps as to how should I do this so that the variable gets populated.
    Thank you.
    Regards,
    TR.

  • How do i test split by value functionality in mesage mapping with multiple

    how do i test split by value functionality in mesage mapping with multiple values ?
    regards,
    venkat

    repeat your source node. in mapping editor you can view queues by right clicking to mapped element.. selecting Display Queues option.. this will show u your values .
    You can also select this Display Queue option for splitByValue option
    for example
    source--->splitByValue>target
    Try viewing your queues to each this step... for splitByValue in display Queue you will see context inserted(grey colour) accodingly

  • Auto Step functionality in Oracle Forms?

    Hello,
    I am testing Captivate for possibility to capture test
    scenarios for testing in Oracle. I know Captivate has the ability
    to create auto steps from Internet Explorer to automatically create
    captions that contain text describing your actions.
    Is this functionality available in Oracle Forms? I have
    tested it many times and Captivate seems to only be able to create
    auto steps from the browser and not create any from Oracle. It
    still captures everything properly, but not the auto steps. We are
    planning on printing these steps for ease of the user, and thus
    would like to have the steps in writing as well.
    This is what the Captivate Help file says about Auto Text:
    "Use the AutoText Captions feature to automatically create captions
    that contain text describing your actions. Customize the
    appearance, size, orientation, position, and display speed of
    captions, including font, size, color, and text formatting. Before
    recording, choose the caption type and language, or create and
    import your own caption style. Adobe Captivate 3 generates caption
    text in many languages."
    Thanks a lot.
    -Zeb

    Welcome to our community, Zeb
    At first, you really bowled me a wobbly with the description
    of "auto step functionality".
    It's been my experience that Captivate's ability to properly
    recognize elements of an application are pretty much hit and miss.
    To be honest, I'm surprised that it does as well as it does. But
    that's with normal desktop applications. Toss in something that's
    browser based (as your situation seems to be) and I'm surprised it
    works at all.
    I'm sort of stabbing in the dark here, as I'm not an
    application programmer. But I'd guess that what is happening is
    Captivate is only sensing an object in the browser and it has no
    idea as to how to interpret the fields and whatnot that make up the
    object.
    I know it's a nice feature when it works. However, I do hope
    you are aware that you aren't limited to what Captivate gives you
    with this. You are totally able to insert your own Text Captions at
    any time on any slide by clicking Insert > Text Caption.
    Cheers... Rick

  • Step function

    I have created a step function using a case structure within a while loop using shift registers. (I'm sure this probably isnt the best way to make a step function but its the only way I know how) Anyways I have two questions regarding my step function.  First: How do I make the function start at 0 each time I start the program? If I stopped the program at 4, it will begin at 4 when I restart the program.  I want it to begin from 0 next time I restart the program, even if I stopped it at 4 the previous run.  Second: The graph of my step function looks like a sawtooth wave rather than a step function because it is only taking one measurement before moving on the the next step.  How do I get it to look more like a step function? I have attached the VI of my step function.  I would also appreciate it if somebody shows me how to make a step function in a more efficient way. I know there is a step function VI but I could not figure out how it works.
    Attachments:
    Step function.vi ‏14 KB

    I found this on the forums a while ago and I am actually using it in a project right now. So, I can't take credit for coming up with this. I definitely suggest doing a search before posting, because often times your question will already be answered. This essentially does a ramp pattern, and uses rounding of decimals to round the numbers off to a flat line if you're between two step values. You can find the ramp pattern VI in the palettes. 
    CLA, LabVIEW Versions 2010-2013
    Attachments:
    stepfunction.png ‏5 KB

  • How to get sysnr value in a Function Module

    Hi all,
    I need to get the sysnr(system number) value of the R/3 system. I execute a RFC function module and need to get the value of sysnr of the system it executes in as a return parameter. Can somebody tell me how to get this value in the function module and return it.
    Thanks and Regards,
    Pratik

    Hello Pratik
    The system number has to be defined in the RFC destination. Thus, select on your local system (where you call the RFC function module) the corresponding RFC destination from table <b>RFCDES</b>. In field RFCDES-RFCOPTIONS you will find a string like this:
    H=<ip address>,S=21,R=N, ...
    S=system number
    Regards
      Uwe

  • After the most recent firefox update I can no longer copy and paste. the fuinction only works in Windows and I HATE usign it. Any suggestions on how to restore the copy and paste function while using Mozilla?

    I have been unable to use the copy and past command while using mozilla. the fucntion does work in windows IE, but I don't want to use windows. How do I restore or fix this function? I varies, works some days and not others.

    See http://kb.mozillazine.org/Clipboard_not_working
    Try to use the keyboard if the buttons on the web page aren't working.
    * Copy: Ctrl+C or Ctrl+Insert
    * Paste: Ctrl+V or Shift+Insert
    * Cut: Ctrl+X or Shift+Delete

  • How to find out what are the functions supported by string class

    Hi,
    Can any one let me know how to find what are all the functions supported by the string class in standard(STL) library on solaris.
    Regards,
    Vignesh

    1. Any C++ textbook that covers the Standard Library will tell you about the standard string class. A good tutorial and reference for the entire Standard Library is "The C++ Standard Library" by Nicolai Josuttis, published by Addison Wesley.
    2. WIth Sun C++, the command
    man -s3C++ basic_string
    provides documentation for the default libCstd version of the Standard Library.
    3. You could look at the <string> header itself. I don't recommend that approach.

  • How to include a new Z Partner Function in Sales order header drop down lis

    Hello SAP experts,
    We have come up with a new requirement to include a custom Partner Function-ZI  in dropdown of SO header-Partners field.
    Current scenario- Presently we are maintaining this partner function  ( Say ZI-  Institute ) in customer master data-Sales area tab.There could be more than one partners under this PF in master data. While entering sales Order, all partners associated are displayed in Header- partners tab, where only SP, SH, BP, PY partner functions are displayed.
    Requirement- 1. Is to  provide Drop down of this ZI-Institute partner function (which I did not find presently) in SO header-partners field and user should be able to provide one of partner function as available in Customer master ( as there are more than one partners that can be set in Customer master data but user need to be able to provide ONE of this ZI partner fucntion in DROPDOWN of PF tab in SO header ).
    2. In addition, How can I link this new partner function in related sales reports against that particulat order and invoice?
    Kindly suggest your valuable inputs as what should be the procedure to get this.
    Many thanks in advance.

    Hi,
    Go to partner determination by VOPA
    OR
    IMG >> SD >> Basic function >> Set up partner functions >> set up partner function for sales document header
    click on partner function and Maintain partner function ZI  and back
    Then select your existing partner determination procedure (Standard is TA) and click on PARTNER FUNCTIONS IN PROCEDURE and add partner function. and make it mandatory so even user forget to mention then at the time of saving system can remind user 
    Then double click on PARTNER DETERMINATION PROCEDURE ASSIGNMENT
    Assign partner procedure to sales document
    Now when you creating sales order go to header partner tab and enter you institute partner number
    For sales order and invoice report
    Execute this report by selecting partner function ZI in VA05N and VF05 respectively
    kapil

  • I have a box of dvds, and i would like to transfer them onto my computer on itunes and then onto my iPad. Does any know how to do this in a step by step process?

    I have a box of dvds, which I would like to transfer to my intunes account on my laptop. I then what to transfer the dvds to my ipad. Does anyone know how to do this in a step by step process?

    If you have an iPod Touch running iOS 4.3.3 or later, you can download the songs in the purchased tab of iTunes on the iPod.
    http://support.apple.com/kb/ht2519

  • How can I use User-Defined Aggregate Functions in Timesten 11? such as ODCI

    Hi
    we are using Timesten 11 version and as per the documentation, it doesn't support User-Defined Aggregate Functions.
    So we are looking for alternatives to do it. Could you please provide your expert voice on this.
    Thanks a lot.
    As the following:
    create or replace type strcat_type as object (
    cat_string varchar2(32767),
    static function ODCIAggregateInitialize(cs_ctx In Out strcat_type) return number,
    member function ODCIAggregateIterate(self In Out strcat_type,value in varchar2) return number,
    member function ODCIAggregateMerge(self In Out strcat_type,ctx2 In Out strcat_type) return number,
    member function ODCIAggregateTerminate(self In Out strcat_type,returnValue Out varchar2,flags in number) return
    number
    How can I use User-Defined Aggregate Functions in Timesten 11? such as ODCIAggregateInitialize ?

    Dear user6258915,
    You absolutely right, TimesTen doesnt support object types (http://docs.oracle.com/cd/E13085_01/doc/timesten.1121/e13076/plsqldiffs.htm) and User-Defined Aggregate Functions.
    Is it crucial for your application? Could you rewrite this functionality by using standart SQL or PL/SQL?
    Best regards,
    Gennady

  • How to use database look up table function in xsl mapping

    Can anybody tell me how to use database look up table function while mapping xsl between 2 nodes.
    I have an XML file coming in and depending on one of XML elements we need to decide which further path to take. But, using this XML element, we need to query database table, get metadata and accordingly take appropriate path. I have written lookup function which returns metadata value.
    Now, the issue is how do I pass the XML element valu as input to look up function? When I tried to drag it to the input node of lookup function, it throws an error like "Maximum number of parameters exceeded"
    Thanks,

    If the lookup table is always going to remain the same (e.g. a character generator or something similar) you can place the values in a 2D array constant on your diagram, with the input value as one column, the equivalent as the other. When you need to perform the lookup you use an index array to return all the values in the "input column", search it using "search 1D array" and use the resulting index number to index the other column's data. If the values may change, then it would probably be best to load an array control with your equivalent values from a file.
    P.M.
    Putnam
    Certified LabVIEW Developer
    Senior Test Engineer
    Currently using LV 6.1-LabVIEW 2012, RT8.5
    LabVIEW Champion

  • How to remove the rule or class function in CS5

    i need to know how to remove the rule or class function in CS5  at the bottom of the screen there are two options for formating HTML and Css when i click the HTML it only allows me to change the bold or italics or link something but when i click CSS it allows me to format how i want the paragraph aligned and the text size and font when i click on lets say changing the font size a box comes up asking me to name a rule so it applies it to everything else i type i want to know how to stop tht like edit everything on my own and if i use CS5 here will it be compatible with CS4 or CS3 at my skool plzz help ive been frustrated with this

    If I use CS5 here will it be compatible with CS4 or CS3 at my skool plzz help ive been frustrated with this
    Code is code.   It doesn't matter which product you use.
    i need to know how to remove the rule or class function in CS5
    You can't.  DW encourages you to use good coding methods, which means using CSS classes and to keep content (HTML) separate from styles (CSS).  For example, if you change font-size on p tags like so:
         p {font-size: 38px}
    Every paragraph will have 38px sized text.
    If you want to apply a special style to just a portion of your text, you must define a CSS class name like so:
    .foo {
    font-size: 38px;
    color: red;
    HTML:
    <p>This is normal paragraph text <span class="foo"> And this is very big and red.</span></p>
    This is normal paragraph text And this is very big and red. 
    Nancy O.
    Alt-Web Design & Publishing
    Web | Graphics | Print | Media  Specialists 
    http://alt-web.com/
    http://twitter.com/altweb
    Message was edited by: Nancy O.  -- unfortunately, this forum doesn't support Raw HTML with inline styles. You'll need to paste my code examples into your DW page to see the effect.

  • How can i do the upload file function using tomcat library??

    how can i do the upload file function using tomcat library??

    Did you read the document for the library?
    If you can't figure it out, why don't you ask the people who provide the library?
    This has nothing to do with JavaMail.

Maybe you are looking for