Looking for advanced books on error/data analysis

[franciobr]

My current level of understanding of error/data analysis and statistics is based on the material I covered from Bevington and Taylor's introductory texts. I need something that covers the subject in greater depth now. My motivation comes from my job which is to come up with a better way of fitting and measuring the "goodness of fit" of a procedure.

Hence, I need a book that covers several fitting techniques, goodness of fit measurements and outlier/anomalies removal (numerical) methods. Also, I want a book that has an applied mathematics mindset, I do not want a book that covers the mathematics in excruciating detail but does not bothers to give a good practical example. Any suggestions?

 

[franciobr]

Any help...pleaase? Think of me as a beauty in distress doomed to never understand advanced statistics unless a hero shows up!=D!

 

[Stephen Tashi]

Any help...pleaase? Think of me as a beauty in distress doomed to never understand advanced statistics unless a hero shows up!=D!

Perhaps if you post a picture...

Otherwise I think you need to define the subject matter of interest more precisely. I don't what "error/data" analysis means. If you want another book that covers the material in Bevington and Taylor, I don't know what's in Bevington and Taylor. As to covering mathematics in excruciating detail - well, this is the math section of the forum.

 

[franciobr]

Ok, I will try to be more specific.

My current level of understanding given by Bevington and Taylor: I know the most famous distribution functions (normal, binomial, poisson), I know the most basic goodness of fit measurements (chi squared) and a very basic outlier removal test (Chauvenet's anomaly removal test) and I understand the least squares fitting technique. Basically, I've read a "introduction to statistics for phisicist" book.

I know that there is a lot more out there as far as statistics inference goes. I want a good source for more advanced: goodness of fit measurements, anomaly removal tests and its relation to several fitting methods. And I don't want a highly theoretical book. I want a book with an applied mindset with tons of examples and tips. I want to understand statistics as a tool, not as my study field which I have to understand every single detail. I posted this here because is the stastics part of the forum, do you believe I should ask for suggestions at another section?

 

[Stephen Tashi]

do you believe I should ask for suggestions at another section?

I don't think it was wrong to post it in this section, but it wouldn't hurt to ask this question in the physics or engineering sections where peple with similar goals may hang out.

My personal slant on statistics is that people who don't study the math in detail don't really know what they are doing. They tend to have msconeptions about fundamental ideas such as "significance" and "confidence". So I don't collect the type of books you describe.

 

[franciobr]

I don't think it was wrong to post it in this section, but it wouldn't hurt to ask this question in the physics or engineering sections where peple with similar goals may hang out.

My personal slant on statistics is that people who don't study the math in detail don't really know what they are doing. They tend to have msconeptions about fundamental ideas such as "significance" and "confidence". So I don't collect the type of books you describe.

Look, I never meant to say "I do not want to study math, I want to skip every basic concept, I just want to magically use it without having any idea of what is going on". I said "I do not want to study only the math/theory, I want to study it in an applied context". I got books in the past that were highly theoretical and I had a hard time applying the theory without examples of applications to be based on.

 

[chiro]

Hey franciobr and welcome to the forums.

You should probably start off by getting a decent book on regression modelling covering a wide range of models from simple linear regression to the more advanced generalized linear mixed models and other abstract general model frameworks.

If you go through these then you will understand how models in a statistical sense are constructed and then you can look for specific concepts related to your problem at hand.

One thing about learning is that learning the lingo and jargon is part of the work: once you know the jargon you'll know what to search for in a specialist field and doing so requires you to get some kind of standard textbook on the subject.