[R] Coefficient of Partial Determination

[David Winsemius]

> On Mar 1, 2017, at 9:57 AM, AbouEl-Makarim Aboueissa <abouelmakarim1962 at gmail.com> wrote:
> 
> Dear All:
> 
> Can the *Coefficient of Partial Determination* in multiple linear
> regression be computed in R? If so, could you please let me know how?
> 
> 
> fullmodel <- lm(Price ~ Size + Lotsize + Bedrooms + Bathrooms)

Partial R^2's should just be a simple calculation of a ratio of sums of squares. After searching on that 'fullmodel" construction I see that a similar problem was posed for college coursework:

houses = read.csv("http://home.cc.umanitoba.ca/~godwinrt/3180/data/houseprice.csv")

(Homework is not considered on-topic for r-help.)

Perhaps you need to learn to search:

sos::findFn("coefficient of partial determination")

... finds one.

sos::findFn("partial R^2")

... finds both that one and another

Code can be found searching the archives:

http://markmail.org/search/?q=list%3Aorg.r-project.r-help+coefficient+of+partial+determination

This had a particularly compact and well commented example:

http://markmail.org/search/?q=list%3Aorg.r-project.r-help+coefficient+of+partial+determination#query:list%3Aorg.r-project.r-help%20coefficient%20of%20partial%20determination+page:1+mid:s7gcefsew5cce46q+state:results

> 
> 
> *I found this in the internet, but I could find the package "rms"*

I'm assuming you meant to type "couldn't", although that is rather surprising since it is a well-established and respected package:

https://cran.r-project.org/web/packages/rms/index.html


> 
> *library(rms)  # will also load Hmisc*
> 
> *fit <- ols(y ~ x1 + x2, data=bf.dat) *
> 
> *plt <- plot(anova(fit), what='partial R2')*

The asterisks would need to be removed but generally you will also need to use a call to `datadist` and `options` when using the rms/Hmisc suite of functions. The book that supports that package is excellent.

Hope this helps;
David.


> 
> *plt*
> 
> 
> 
> Here is part of the data as an example:
> ---------------------------------------------------
> 
> Taxes Bedrooms Bathrooms   Price Size Lotsize
>   296        3         3  795000 2371    5850
>   242        4         3  399000 2818    4000
>   242        4         3  545000 3032    3060
>   222        4         4  909000 3540    6650
>   222        3         1  109900 1249    6360
>   222        3         3  324900 1800    4160
>   311        4         2  192900 1603    3880
>   311        3         2  215000 1450    4160
>   311        4         3  999000 3360    4800
>   311        3         2  319000 1323    5500
>   311        3         2  350000 1750    7200
>   311        3         2  249000 1400    3000
>   311        2         2  299000 1257    1700
>   307        3         2  235900 1400    2880
>   307        3         2  348000 1600    3600
>   307        4         3  314000 1794    3185
>   307        4         2  399000 1850    3300
>   307        3         3  599000 2950    5200
>   307        3         2  299000 1719    3450
>   307        3         3  425000 1472    3986
>   307        4         3 1100000 4168    4785
>   307        3         3 1500000 3880    4510
>   307        2         1  110000 1000    4000
>   307        3         2  200000 1139    3934
>   307        3         1  134900 1080    4960
>   307        4         3  250000 2000    3000
>   307        3         4  950000 1920    3800
>   307        4         2  239950 1348    4960
>   307        3         2  170000 1280    3000
>   307        3         2  285000 2400    4500
>   307        3         3  279000 1700    3500
>   307        3         2  219000 1600    3500
>   307        3         2  155000 1050    4000
>   307        3         2  389000 1415    4500
>   307        3         1  340000 1110    6360
>   279        2         1   95000  797    4500
>   279        2         2  140000 1100    4032
>   279        3         3 1100000 2602    5170
>   279        4         3  360000 2351    5400
>   252        3         3  415000 1350    3150
>   252        4         2  250000 1206    3745
>   233        3         3  559000 2628    4520
>   233        3         3  525000 2365    4640
>   233        3         3  779000 2990    8580
>   233        3         2  595000 1750    2000
>   233        4         5 1150000 5500    2160
>   233        3         2  550000 1852    3040
>   233        3         2  500000 2100    3090
>   233        4         3  279000 2580    4960
>   233        4         2  375000 1963    3350
>   243        3         3  330000 1900    5300
>   243        3         3  199000 1450    4100
>   243        2         2  165000 1000    9166
>   243        4         3 1399000 6500    4040
>   469        3         2  255000 1218    3630
>   226        2         2  325000  893    3620
> 
> 
> 
> Thank you very much for your help and support
> 
> abou
> ______________________
> AbouEl-Makarim Aboueissa, PhD
> University of Southern Maine
> Department of Mathematics and Statistics
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

David Winsemius
Alameda, CA, USA