Semester 2 Chatter Thread (2009) (2 Viewers)

[wixxy2348]

i dropped it in favour of:

philosophy of mind



I find it somewhat iRONic that a faculty which promoted equality, freedom of speech, tolerance yada yada yada has none of that for anything which isn't in the same viewpoint as their own.

all im asking for is a lack of bias mang

MALFOY WAS RIGHT

Lol.

 

[Silver Persian]

Comrades, refresh me on what I've forgotten from first year statistics

I have two straight lines in an Excel graph, one a mathematical model (fitted via regression analysis) and the actual raw data. How do I go about expressing the error between these two lines? Do I take the standard deviation of each point, and then average it? Or is there an easier way?

HALP

I would recommend proving that the raw data is a meaningless, culturally biased expression of the data gatherer's conceptual framework. Therefore, any statistical inference gained from it is null, and there is no point doing aforementioned equations.

Was this helpful?

 

[jb_nc]

Comrades, refresh me on what I've forgotten from first year statistics

I have two straight lines in an Excel graph, one a mathematical model (fitted via regression analysis) and the actual raw data. How do I go about expressing the error between these two lines? Do I take the standard deviation of each point, and then average it? Or is there an easier way?

HALP

Correlation co-efficient, R?

 

[Ben Netanyahu]

yes. correlation! v. good. that's not provided by the excel regression analysis so i suppose i'll have to punch the numbers in by hand or something.

thanks jb_nc

SP: eat poo

 

[jb_nc]

No, it is. It's in trendline properties under (or above) "show equation". This might help:

Excel Tutorial on Linear Regression

 

[Ben Netanyahu]

actually, if i remember correctly, doesnt correlation measure only a single data set? i want to measure the correlation between two data sets. eg two linear equations

 

[jb_nc]

Well, if the equations are linear, the error is going to be constant. Pick two points, evaluate for f(x) and then compute the percentage difference?

 

[Ben Netanyahu]

screenshot shows basically what i'm trying to do.

two linearised models of adsorption (blue one is the linear one i believe). need to determine which is best through determining the error of each, i'm supposing. i tried to measure the average of the deviation from each data point and then compare them until i realised that each was on a separate scale and that my calculations there had become meaningless.

 

[jb_nc]

I'm not sure. I think you mean something like this: Error Analysis

??

h8 error analysis m8

 

[Ben Netanyahu]

yeah same

good thing my lab partner is good at it, and physical chemistry in general

i'm making it up by typing out the non mathsy parts of the report right now.

 

[withoutaface]

Fuck Java is terrible. I might be the only person who finds it more cumbersome to program in than C++, but the lack of generic programming really shits me.

EDIT: Also all these third party APIs eat balls.

EDIT2:

Not it fucking doesn't. If you'd implemented your interface intelligently instead of being an absolute pack of arseholes I'd have a linkedhashmap to iterate through, but your descriptions don't match your fucking implementation and now I have to hack my way around it.

 

[Ben Netanyahu]

all the people i know who program, all two of them, quite enjoy java and use it often

i know html

go me

 

[withoutaface]

Java is programming for 12 year olds. They refuse to let you use more advanced functionality because they don't think you're intelligent enough to use it.

 

[kokodamonkey]

hmm. Forgot what i was going to say! shame.

Im actually starting to work a bit now... so working 11-5 tomorrow means i have to finish ma essay tomorrow night! and then do all the homework due monday.. bah!!!!

 

[kokodamonkey]

goodnight nebz

 

[imsopostmodern]

head is killing me from crying so much

 

[wixxy2348]

aww

 

[lala2]

Zomg I've finally perfected my resume. I've been experimenting with different fonts and layouts for a while already, and I've got one I think I like. I think I'll leave selection criteria for tomorrow--it's just so much easier to think up answers for them when your CV inspires you to think

 

[kokodamonkey]

post it.

 

[scarybunny]

Grr I saved some journal articles that it took me yonks to find on my stick, and they're both corrupted.

BACK TO ERIC. Gay.