iSixSigma

Best fit not regression

Six Sigma – iSixSigma Forums Old Forums General Best fit not regression

This topic contains 8 replies, has 5 voices, and was last updated by  Thomas Trible 14 years, 8 months ago.

Viewing 9 posts - 1 through 9 (of 9 total)
  • Author
    Posts
  • #37276

    AT
    Participant

    I have regression fit for the given data but that is not what i am looking for I am not interested in regression analysis thats not my aim I want to plot a curve which can give me a best fit line for the data I have and I need help with it.
    Thanks.

    0
    #109421

    Doodnauth
    Participant

    Interesting. What do you see as the difference?

    0
    #109438

    AT
    Participant

    My Understanding says Regression does give a equation giving out a best fit line  but what I am looking at is Just draw a bar chart and to fit the frequency to the best possible probabilty distribution depending on the data I have, i hope you understand what i am talking about.
    Please let me know if we could do it by anyways.
    thanks
    ramesh!

    0
    #109440

    AT
    Participant

    I would really like to know what the difference is if theres any and want to find out what and when either will be used.

    0
    #109442

    RubberDude
    Member

    Sounds like the data is not regression data.  If not, what type data is it?  What are you trying to “best fit” to?  I think we need more details in order to help you.

    0
    #109456

    Pepsi
    Participant

    If you have minitab, “show descriptive statistics” of your data, and show the “graphical summary”. A histogram will be displayed, with your probability distribution overlayed.
    I beleive this probability distribution is what you are referring to as your “best fit line”.

    0
    #109463

    AT
    Participant

    I have data in Two columns and one column is 24hours in a day and the 2nd column is Total No. of arrivals in every Hour of the day for all 24 hours Data, using this I plotted a Bar chart and i got the frequency distribution graph for Hours Vs No. arrival and after I draw this I find out the flow patter, I mean the arrival can be described well by what type of probability distribution to find the best fit line I amtrying to do and I want the equation of that line and I dont want to Regression analaysis but still I want a Best Fit line.
    Thanks.

    0
    #109524

    RubberDude
    Member

    In a simple MS Excel spreadsheet you can produce a simple bar chart and then right click one of the bars on the chart.  One option on the menu that pops up will be “add trendline.”  There are several trendline options available.  I’m thinking you will want a polynomial type.  Try that and see if that’s what you are looking for.
    RubberDude – Certified Grand Master Holiday Inn Express Black Belt

    0
    #109530

    Thomas Trible
    Member

    Ramesh:
    Sounds like you have time series data.  You didn’t say what your data analysis objective was, but let’s say for the moment that you want to understand something that’s significant about the data.
    I recommend that you plot your time series data first on a run chart, and then for added interpretation, on a control chart.  Applying run chart rules will allow you determine if there is an actual trend, or not.  One potential problem with applying a trend line to raw data, or using linear regression analysis to do so, is that you are making a statement of significance. Taking the argument a bit further, you are saying that the observed “trend” that you’ve just “discovered” on your data is due to special cause variation, when in fact the variation may be only random.
    The run chart will enable you to determine if anything “significant” is happening with the data, so that you can begin looking for those special causes, if there are any.
    TC Trible

    0
Viewing 9 posts - 1 through 9 (of 9 total)

The forum ‘General’ is closed to new topics and replies.