iSixSigma

Informal Survey on Advanced Tools

Six Sigma – iSixSigma Forums Old Forums General Informal Survey on Advanced Tools

Viewing 16 posts - 1 through 16 (of 16 total)
  • Author
    Posts
  • #67744

    Kerri
    Participant

    I too would like to know the answer. Can I request that they be posted to the forum for me (and others) to read and learn also? I’m concerned that if users just put “respond to me personally”, that this discussion forum will not be a “discussion forum” and instead will turn into a bulletin board. There is a difference.
    Thanks,Kerri

    0
    #67745

    John Noguera
    Participant

    Kerri,
    Good point.
    Please respond on the board rather than by e-mail.
    Thanks,
    John
     

    0
    #67746

    Anonymous
    Participant

    I’m a Ph.D. statistician and Black Belt. In my dealings with about 5 projects (so far), there has been the request for, and the use of, several advances techniques.
    We have done Poisson regression, multivariate SPC, ROC analysis, bootstrapping and some advanced robust design techniques.
    We do much more outside of Six Sigma (logistic regression and neural nets among others).
     

    0
    #67747

    melvin
    Participant

    Our manufacturing firm has internalized their Six Sigma teachings and within the teachings, none of these topics are present.  Now, we as Master Black Belts recognize the occasional need for tools such as response surface modeling, cluster analysis, etc. and when we do, we discuss these topics individually with the black belts.
    In a nutshell, the use of the tools is limited and instruction on their use is on an as needed basis.

    0
    #27572

    John Noguera
    Participant

    I am interested in getting a sense on the actual use of the more advanced statistical tools by Six Sigma BBs or MBBs.
    Have any of you applied any of the following tools with successful outcome (other than just experimenting to learn the tool):
    Multivariate Tools (Principal Components Analysis, Factor Analysis, Cluster Analysis, Discriminant Analysis, Partial Least Squares)
    Multivariate SPC
    Variance Components SPC
    EWMA
    Time Series Analysis (ARIMA, Exponential Smoothing)
    Classification and Regression Trees (CART)
    Logistic Regression
    Bootstrap Methods
    Neural Networks
    Any response would be greatly appreciated.  My e-mail address is [email protected].
     
     

    0
    #67752

    Ken Myers
    Participant

    John,Our organization makes extensive use of two tools that are typically considered more on the advanced side. One tool is called Change Point Analysis(CPA). We use CPA extensively in our problem solving efforts to identify significant trendwise changes in time series data that are normally not identified using conventional SPC methods. CPA combines the use of CuSum and bootstraping to perform this analysis.The second tool is Variation Transmission Analysis(VTA). We typically use VTA after conducting a Response Surface Study to establish tolerances for the key inputs of the process.Ken

    0
    #67755

    Leung
    Participant

    I recently used both principle component analysis and cluster analysis, but not in support of any six sigma projects; some design engineers brought by some data and wanted me to play around with it (the data)It proved useful, so they keep coming back with other data.
    I am currently working on a project in which I will use Multivariate SPC; this is for my own curosity — it could be used later on when the program goes into production. The customer wants about 15 to 20 control charts; I think I may be able to greatly reduce that number or at least get them thinking abou it.
    Finally, I frequently use time series analysis in conjunction with my six sigma projects. There is a nice intuiative touch to it that team members find useful.
     A note of caution — I have access to a real statitistician who keeps me on the straight and narrow.

    0
    #67765

    Ken K.
    Participant

    Don’t forget Reliability Analysis – parametric and nonparametric analysis of censored data (usually lifetime data). Very important when analyzing field failure data, or censored laboratory data (common example – pull tests where the fixture or part fails before the adhesive).

    0
    #67870

    John Noguera
    Participant

    To Anonymous, Bob, Ken Myers, Ben, Ken K, and Nathan:
    Thank you for responding to my inquiry on the use of advanced tools. 
    Here is my summary of the tools mentioned:
    Poisson regressionMultivariate SPC (2)Receiver Operating Characteristic (ROC) analysisBootstrapping (2)Advanced robust design techniquesLogistic regressionNeural netsResponse surface modeling (2)Cluster analysis (2)Change Point Analysis (CPA-combines the use of CuSum and bootstraping)Variation Transmission Analysis (VTA – tolerancing)Principle Component AnalysisTime Series AnalysisReliability Analysis – parametric and nonparametric analysis of censored dataMonte Carlo simulation,  Sensitivity, Tornado (Crystal Ball)Most of these topics (with the exception of RSM) are typically not covered in a Black Belt curriculum.   Do you folks think that the availability of supplementary training in these topics would benefit Black Belts and be in demand?

    0
    #67899

    Ronald
    Participant

    Here we go again, assuming BB means statistician. How about a BB being a leader. In my years as a BB, MBB and now Director I haven’t seen many of those tools used, BUT my experience is with the non-mfg side of the house. I would rather my BB’s know how to facilitate and lead high performance teams than advanced stats.

    0
    #67901

    John Noguera
    Participant

    Does being a leader of a high performance team necessitate the use of sub-optimal tools?  The purpose of the informal survey was to get an indicator of those advanced tools that really do assist BBs and MBBs.

    0
    #67902

    John Noguera
    Participant

    Another point: Virtually any advanced tool can be utilized successfully without having to delve heavily into the theory.   That’s why we use statistical software!

    0
    #67955

    Scott
    Member

    Most definitely

    0
    #67957

    Ronald
    Participant

    This is why Six Sigma is such a struggle to implement in the non-manufacturing arena, stats guys are involved. Believe me, i know stats, and I know the software. Statistics is a powerful tool, However, the emphasis need not be on Stats as it always seems to be. Six Sigma is a decision making methodology that has proven itself over time. Most of the advanced tools are of limited need in the non-mfg environment, Yet stats guys try to force there advanced tools in there and then wonder why a VP of sales doesn’t want to talk to them.
    IF you want to do a survey on advanced tools…then do one. In fact, i have one already completed from both a student and instructor (MBB) level if you would like the feedback. You may be surprised at the value stats plays in all this.
     

    0
    #67960

    John Noguera
    Participant

    Lee,
    Thank you for your feedback. Yes, I would be very interested in the surveys you mentioned.  Please e-mail me at [email protected], or post on this board.
    I would have to agree with you that many of the advanced tools are NOT applicable in the transactional arena, however SOME from the summary list are very useful, including Process Simulation (to which all the DOE tools can be applied), Logistic Regression, Neural Nets and Time Series Analysis.
    I am also a big proponent of Basic Tools such as Process Mapping, Pareto, Cause-and-Effect, etc…  The advanced tools only apply after you have cleared the deck of the low hanging fruit, and you are moving on trying to resolve chronic problems that may be due to hidden interactions.
     

    0
    #67965

    Kim Niles
    Participant

    I spent a lot of time going through all the books I could find on this subject to determine that reason #8 is in common with those authors.  See “What Makes Six Sigma Work” at: https://www.isixsigma.com/library/content/c010723a.asp ; Reason #8: “Quality tools that never get used are thrown out. If we don’t need them, why spend time learning how to use them”.
    However, I personally enjoy this discussion and feel that one can never learn too much about tools and techniques, as long as it’s cost effective to do so (that’s the kicker).     
    KN
     

    0
Viewing 16 posts - 1 through 16 (of 16 total)

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