At 6:53 the MTTF is calculated to be 8103. At 7:05 the template for the same parameters gives the mean as 5101. I don't understand why they are different?
Hi Hemant, you are very good and thank you for your classes, but please in simple words explain why these different distributions are important and how to be used in reliability analyses. We tend to forget what the objective is sometimes, I am sure it is buried in your tutorials but it will be nice to have a clear simple explanation. I have done a lot of analyses for EPSA, WCA and SEEA for NASA but using the distributions functions, I think is for estimating the future outcome per present or past knowledge of yields. So are we making models for predicting failure rates in future per sampling of the population and fitting the sample date to a proper distribution to guess the future outcome? Please forgive me for my lack of knowledge. Regards, Ash Tehrani
The different distributions are required to fit data better. One way to decide is to use maximum likelihood estimate (MLE) reach is a mathematical procedure using calculus. Another procedure is to find differences between observed and fitted values with the selected distribution. The one that minimizes this is selected. Statistically it is called goodness of fit test. It is like finding which size shirt fits one! No distribution would exactly fit the data. Hope this helps.
Greetings Mr. Urdhwareshe , Firstly, thank you for the video. I was in need of simple instructions to conduct simple data distribution plotting in Minitab. in that regard, i do have a question if you don't mind! I am trying to calculate Sauter Mean using log-normally distributed data with the following parameters: u = -2.22 um σ = 0.43 particles diameter range : 0.5 um - 300 um So, my question is how to correctly substitute the values with the mean being negative?
Hello Hamzah Naser! Appreciate your feedback and keen interest! I don't think you can use a negative value in the lognormal. Perhaps the only way that may be possible is to add a constant value to all data to make it positive. I don't know whether this helps you! Good luck..Hemant
@@instituteofqualityandrelia7902 Thank you for the quick reply, Mr. Hemant. I believe it is possible to have a negative mean for log-normally distributed data (or at least that's what i found out). I have tried removing the negative sign - a multiplication of (-1) as a constant if that's what you were referring to - but i am still unsure if it's a valid assumption or not! Would an integration of that range with the negative sign removed = a value close to 1 indicate a valid assumption? Once again, thank you very much for the quick reply.
Sir, a very good and informative video. However, I can not understand the difference between MU Dash (the log mean) and XBar (in the standard deviation formula). Would you be kind enough to tell me what the difference is. Thank you.
Thanks. I think I should have used 'x-bar dash' instead of x-bar to indicate that that it is a sample average of logarithms. x-bar indicates sample average while as mu-dash is a population average. x-b.
Thank you so much. I have one question. How can I do largest extreme value based-on reliability analysis by Minitab?. Why I ask this question, because in Minitab software only exist smallest extreme value in "Reliability/Survival" section. I tried to understand the solution, but I couldn't to figure it out. Or is it possible for you to make a video about the largest extreme value by Minitab and share?. I will very appreciate for your answer.
Thanks for your quetion and keen interest. I have personally not used extreme value distributions. But as I understand from the Minitab support topic, "The largest extreme value (LEV) distribution and the smallest extreme value (LEV) distribution are closely related. For example if X has a largest extreme value distribution, then −X has a smallest extreme value distribution, and vice versa." If you see the equations of the largest and smallest extreme value distributions, (www.weibull.com/hotwire/issue128/relbasics128.htm) you will realize the same thing. So perhaps (my guess) you need to multiply the data by -1 and fit SEV. (link support.minitab.com/en-us/minitab/18/help-and-how-to/probability-distributions-and-random-data/supporting-topics/distributions/smallest-and-largest-extreme-value-distributions/) . But I suggest askiing Minitab support and getting confirmation of this.
I have not seen any requirement to calculate mode of lognormal distribution. However, formula is available as exp(logmean -(log Std Deviation)square). Any specific reason for the question?
I don't know what magic this man uses, but I've listened to my teacher and other videos in my native language (which is french), and didn't grasp the idea that you had to apply the logarithm ON THE DATA, since it is this new value that follows a normal distribution, like ''translating'' words to another language to work and ''retranslating'' the results of whatever calculation within the normal distribution, to the original language if you want to work again outside a normal distribution... I swear it has something to do with the indian accent...
Very good explanation sir. Please keep making more these. Thank you!
Thank you Soumya. Please subscribe, if you have not so far. Currently I have released 60 videos. Intend to upload more regularly!
Great video! you explained everything really well
Thank you!
Finally I happy for your work sir
Good to know Santhosh! More videos in pipeline! I was making many music videos for some time and technical videos got delayed as a result!
2:56 in the standard deviation formula, it should be x bar or x' bar?
You are correct. It should have been x' bar. Apologise for the oversight.
Very well explained
I am glad it was helpful! Appreciate your feedback!
At 6:53 the MTTF is calculated to be 8103. At 7:05 the template for the same parameters gives the mean as 5101. I don't understand why they are different?
Thanks for your observation. Will check and get back.
Many thanks
Hi Sir, I have been waiting for your videos.
So nice of you
Really helpful! Thank you for this video
Glad it was helpful!
Great job Sir.
Thank you.
Do you share the Excel Template?
Which template are you refering to?
Hi Hemant, you are very good and thank you for your classes, but please in simple words explain why these different distributions are important and how to be used in reliability analyses. We tend to forget what the objective is sometimes, I am sure it is buried in your tutorials but it will be nice to have a clear simple explanation.
I have done a lot of analyses for EPSA, WCA and SEEA for NASA but using the distributions functions, I think is for estimating the future outcome per present or past knowledge of yields.
So are we making models for predicting failure rates in future per sampling of the population and fitting the sample date to a proper distribution to guess the future outcome?
Please forgive me for my lack of knowledge.
Regards,
Ash Tehrani
The different distributions are required to fit data better. One way to decide is to use maximum likelihood estimate (MLE) reach is a mathematical procedure using calculus. Another procedure is to find differences between observed and fitted values with the selected distribution. The one that minimizes this is selected. Statistically it is called goodness of fit test.
It is like finding which size shirt fits one! No distribution would exactly fit the data. Hope this helps.
Greetings Mr. Urdhwareshe ,
Firstly, thank you for the video. I was in need of simple instructions to conduct simple data distribution plotting in Minitab. in that regard, i do have a question if you don't mind!
I am trying to calculate Sauter Mean using log-normally distributed data with the following parameters:
u = -2.22 um
σ = 0.43
particles diameter range : 0.5 um - 300 um
So, my question is how to correctly substitute the values with the mean being negative?
Hello Hamzah Naser! Appreciate your feedback and keen interest!
I don't think you can use a negative value in the lognormal. Perhaps the only way that may be possible is to add a constant value to all data to make it positive. I don't know whether this helps you!
Good luck..Hemant
@@instituteofqualityandrelia7902 Thank you for the quick reply, Mr. Hemant.
I believe it is possible to have a negative mean for log-normally distributed data (or at least that's what i found out).
I have tried removing the negative sign - a multiplication of (-1) as a constant if that's what you were referring to - but i am still unsure if it's a valid assumption or not!
Would an integration of that range with the negative sign removed = a value close to 1 indicate a valid assumption?
Once again, thank you very much for the quick reply.
Sir, a very good and informative video. However, I can not understand the difference between MU Dash (the log mean) and XBar (in the standard deviation formula). Would you be kind enough to tell me what the difference is.
Thank you.
Thanks. I think I should have used 'x-bar dash' instead of x-bar to indicate that that it is a sample average of logarithms. x-bar indicates sample average while as mu-dash is a population average. x-b.
Thank you so much. I have one question. How can I do largest extreme value based-on reliability analysis by Minitab?. Why I ask this question, because in Minitab software only exist smallest extreme value in "Reliability/Survival" section. I tried to understand the solution, but I couldn't to figure it out. Or is it possible for you to make a video about the largest extreme value by Minitab and share?. I will very appreciate for your answer.
Thanks for your quetion and keen interest. I have personally not used extreme value distributions. But as I understand from the Minitab support topic, "The largest extreme value (LEV) distribution and the smallest extreme value (LEV) distribution are closely related. For example if X has a largest extreme value distribution, then −X has a smallest extreme value distribution, and vice versa."
If you see the equations of the largest and smallest extreme value distributions, (www.weibull.com/hotwire/issue128/relbasics128.htm) you will realize the same thing. So perhaps (my guess) you need to multiply the data by -1 and fit SEV. (link support.minitab.com/en-us/minitab/18/help-and-how-to/probability-distributions-and-random-data/supporting-topics/distributions/smallest-and-largest-extreme-value-distributions/) . But I suggest askiing Minitab support and getting confirmation of this.
Sir, could you explain the mode of log normal distribution
I have not seen any requirement to calculate mode of lognormal distribution. However, formula is available as exp(logmean -(log Std Deviation)square). Any specific reason for the question?
I don't know what magic this man uses, but I've listened to my teacher and other videos in my native language (which is french), and didn't grasp the idea that you had to apply the logarithm ON THE DATA, since it is this new value that follows a normal distribution, like ''translating'' words to another language to work and ''retranslating'' the results of whatever calculation within the normal distribution, to the original language if you want to work again outside a normal distribution...
I swear it has something to do with the indian accent...
Thank you! Appreciate your feedback!