[R] Difficulty with 'loess' function

[Bert Gunter]

Well, you might start by reading the Help file for loess and
specifying the model via a formula, as described there.

The docs say your input is coerced to a formula -- but obviously not
the one you want. Specify it explicitly.

-- Bert

On Fri, Mar 18, 2011 at 8:34 AM, armstrwa <William.Armstrong at noaa.gov> wrote:
> Hi,
>
> I am trying to create a loess smooth from hydrologic data.  My goal is to
> create a smooth line that describes discharge at a certain point in time.  I
> have done this using the 'lowess' function and had no problem, but I'm
> having some difficulty with loess.  I am inputting the date ('date') and
> discharge ('q') values using the 'scan' function, then inputting the
> following:
>
>> dateq<-data.frame(date,q)
>> dateq_smooth<-loess(dateq,span=.5)
>> dateq_smooth_pred<-predict(dateq_smooth,date)
>> dateq_smooth_pred
>  [1] 1959.164 1959.516 1963.379 1967.121 1963.781 1970.601 1958.056 1948.846
> 1954.247
> [10] 1954.718 1969.460 1960.974 1949.845 1962.436 1968.374 1954.247 1956.135
> 1948.852
> [19] 1948.846 1957.893 1963.055 1955.327 1961.453 1960.405 1956.135 1963.065
> 1959.358
> [28] 1954.478 1954.415 1964.323 1963.255 1953.139 1966.048 1959.353 1961.817
> 1959.894
> [37] 1958.251 1964.914 1960.157 1964.656 1964.164 1949.873 1965.500 1968.323
> 1962.196
> [46] 1961.656 1954.977 1969.196 1962.318 1952.576 1962.655 1968.031 1963.445
> 1967.856
> [55] 1961.323 1949.032 1954.687 1955.613 1982.081 1960.736 1963.466 1974.468
> 1970.164
> [64] 1954.212 1957.928 1957.893 1961.985 1966.180 1963.283 1963.379 1964.387
> 1970.017
> [73] 1959.353 1969.925 1963.111 1962.667 1963.417 1964.029 1965.241 1964.323
> 1964.405
> [82] 1963.036 1981.359 1968.146 1968.374 1970.050 1963.089 1959.667 1966.129
> 1958.081
> [91] 1964.029 1968.374 1967.856 1955.758 1954.619 1965.227
>
> This output is not what I'm looking for.  It seems like maybe it's smoothing
> the date data instead of discharge?
>
> I tried messing around with other ways of inputting the data and I was
> getting errors about the input being type "text" and "double" or something
> like that.
>
> I read up on:
> http://research.stowers-institute.org/efg/R/Statistics/loess.htm
> http://research.stowers-institute.org/efg/R/Statistics/loess.htm   but I was
> unable to figure out what I am doing wrong.
>
> Does anyone have any suggestions to offer?
>
> Thanks.
>
> Billy
>
> --
> View this message in context: http://r.789695.n4.nabble.com/Difficulty-with-loess-function-tp3387537p3387537.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org mailing list
> 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.
>



-- 
Bert Gunter
Genentech Nonclinical Biostatistics