You are right Jeff, that was a mistake, I was focusing on the square root and made the mistake of talking about taking the square root instead of raising to the 2nd power.
This is the example I was following ( https://www.youtube.com/watch?v=SaQgA6V8UA4). Of course, I tried fitting the nnet model to my own data, to see what kind of results I'd get (the data that I used, I provided in the very first e-mail). The question I was asking is why do I get a bunch of 1s for the predictions, given that the expected results would have to be somewhere close to the latest observations. The code and the data from the example I was following is provided in the youtube link above. Paul <https://www.youtube.com/watch?v=SaQgA6V8UA4> El mié., 2 sept. 2020 a las 10:01, Jeff Newmiller (<jdnew...@dcn.davis.ca.us>) escribió: > Why would you expect raising y_pred to the 0.5 to "backtransform" a model > sqrt(y)~x? Wouldn't you raise to the 2? > > Why would you "backtransform" x in such a model if it were never > transformed in the first place? Dr Maechler did not suggest that. > > And why are you mentioning some random unspecified video on Youtube? That > does not enlighten anyone here, apparently including you. Please reference > package documentation, and/or reproduce the analysis discussed in that > video to provide a contrasting (or supporting) point with the example you > gave. > > > On September 2, 2020 7:21:58 AM PDT, Paul Bernal <paulberna...@gmail.com> > wrote: > >Dear Dr. Martin and Dr. Peter, > > > >Hope you are doing well. Thank you for your kind feedback. I also tried > >fitting the nnet using y ~ x, but the model kept on generating odd > >predictions. If I understand correctly, from what Dr. Martin said, it > >would > >be a good idea to try modeling sqrt(y) ~ x and then backtransform > >raising > >both y and x to 0.5? > > > >I was looking at a video where the guy modeled count data without doing > >any > >kind of transformation and didn't get odd results, which is rather > >extrange. > > > >Cheers, > > > >Paul > > > > > > > >El mié., 2 sept. 2020 a las 2:37, Martin Maechler (< > >maech...@stat.math.ethz.ch>) escribió: > > > >> >>>>> peter dalgaard > >> >>>>> on Wed, 2 Sep 2020 08:41:09 +0200 writes: > >> > >> > Generically, nnet(a$y ~ a$x, a ...) should be nnet(y ~ x, > >> > data=a, ...) otherwise predict will go looking for a$x, no > >> > matter what is in xnew. > >> > >> > But more importantly, nnet() is a _classifier_, > >> > so the LHS should be a class, not a numeric variable. > >> > >> > -pd > >> > >> Well, nnet() can be used for both classification *and* regression, > >> which is quite clear from the MASS book, but indeed, not from > >> its help page, which indeed mentions one formula 'class ~ ...' > >> and then only has classification examples. > >> > >> So, indeed, the ?nnet help page could improved. > >> > >> In his case, y are counts, so John Tukey's good old > >> "first aid transformation" principle would suggest to model > >> > >> sqrt(y) ~ .. in a *regression* model which nnet() can do. > >> > >> Martin Maechler > >> ETH Zurich and R Core team > >> > >> > >> > >> >> On 1 Sep 2020, at 22:19 , Paul Bernal > >> >> <paulberna...@gmail.com> wrote: > >> >> > >> >> Dear friends, > >> >> > >> >> Hope you are all doing well. I am currently using R > >> >> version 4.0.2 and working with the nnet package. > >> >> > >> >> My dataframe consists of three columns, FECHA which is > >> >> the date, x, which is a sequence from 1 to 159, and y, > >> >> which is the number of covid cases (I am also providing > >> >> the dput for this data frame below). > >> >> > >> >> I tried fitting a neural net model using the following > >> >> code: > >> >> > >> >> xnew = 1:159 Fit <- nnet(a$y ~ a$x, a, size = 5, maxit = > >> >> 1000, lineout = T, decay = 0.001) > >> >> > >> >> Finally, I attempted to generate predictions with the > >> >> following code: > >> >> > >> >> predictions <- predict(Fit, newdata = list(x = xnew), > >> >> type = "raw") > >> >> > >> >> But obtained extremely odd results: As you can see, > >> >> instead of obtaining numbers, more or less in the range > >> >> of the last observations of a$y, I end up getting a bunch > >> >> of 1s, which doesn´t make any sense (if anyone could help > >> >> me understand what could be causing this): > >> >> dput(predictions) structure(c(1, 1, 1, 1, 1, 1, 1, 1, 1, > >> >> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, > >> >> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, > >> >> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, > >> >> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, > >> >> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, > >> >> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, > >> >> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, > >> >> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1), .Dim > >> >> = c(159L, 1L), .Dimnames = list(c("1", "2", "3", "4", > >> >> "5", "6", "7", "8", "9", "10", "11", "12", "13", "14", > >> >> "15", "16", "17", "18", "19", "20", "21", "22", "23", > >> >> "24", "25", "26", "27", "28", "29", "30", "31", "32", > >> >> "33", "34", "35", "36", "37", "38", "39", "40", "41", > >> >> "42", "43", "44", "45", "46", "47", "48", "49", "50", > >> >> "51", "52", "53", "54", "55", "56", "57", "58", "59", > >> >> "60", "61", "62", "63", "64", "65", "66", "67", "68", > >> >> "69", "70", "71", "72", "73", "74", "75", "76", "77", > >> >> "78", "79", "80", "81", "82", "83", "84", "85", "86", > >> >> "87", "88", "89", "90", "91", "92", "93", "94", "95", > >> >> "96", "97", "98", "99", "100", "101", "102", "103", > >> >> "104", "105", "106", "107", "108", "109", "110", "111", > >> >> "112", "113", "114", "115", "116", "117", "118", "119", > >> >> "120", "121", "122", "123", "124", "125", "126", "127", > >> >> "128", "129", "130", "131", "132", "133", "134", "135", > >> >> "136", "137", "138", "139", "140", "141", "142", "143", > >> >> "144", "145", "146", "147", "148", "149", "150", "151", > >> >> "152", "153", "154", "155", "156", "157", "158", "159"), > >> >> NULL)) > >> >> > >> >> head(a) FECHA x y 1 2020-03-09 1 1 2 2020-03-10 2 8 3 > >> >> 2020-03-11 3 14 4 2020-03-12 4 27 5 2020-03-13 5 36 6 > >> >> 2020-03-14 6 43 > >> >> > >> >> dput(a) structure(list(FECHA = structure(c(18330, 18331, > >> >> 18332, 18333, 18334, 18335, 18336, 18337, 18338, 18339, > >> >> 18340, 18341, 18342, 18343, 18344, 18345, 18346, 18347, > >> >> 18348, 18349, 18350, 18351, 18352, 18353, 18354, 18355, > >> >> 18356, 18357, 18358, 18359, 18360, 18361, 18362, 18363, > >> >> 18364, 18365, 18366, 18367, 18368, 18369, 18370, 18371, > >> >> 18372, 18373, 18374, 18375, 18376, 18377, 18378, 18379, > >> >> 18380, 18381, 18382, 18383, 18384, 18385, 18386, 18387, > >> >> 18388, 18389, 18390, 18391, 18392, 18393, 18394, 18395, > >> >> 18396, 18397, 18398, 18399, 18400, 18401, 18402, 18403, > >> >> 18404, 18405, 18406, 18407, 18408, 18409, 18410, 18411, > >> >> 18412, 18413, 18414, 18415, 18416, 18417, 18418, 18419, > >> >> 18420, 18421, 18422, 18423, 18424, 18425, 18426, 18427, > >> >> 18428, 18429, 18430, 18431, 18432, 18433, 18434, 18435, > >> >> 18436, 18437, 18438, 18439, 18440, 18441, 18442, 18443, > >> >> 18444, 18445, 18446, 18447, 18448, 18449, 18450, 18451, > >> >> 18452, 18453, 18454, 18455, 18456, 18457, 18458, 18459, > >> >> 18460, 18461, 18462, 18463, 18464, 18465, 18466, 18467, > >> >> 18468, 18469, 18470, 18471, 18472, 18473, 18474, 18475, > >> >> 18476, 18477, 18478, 18479, 18480, 18481, 18482, 18483, > >> >> 18484, 18485, 18486, 18487, 18488), class = "Date"), x = > >> >> 1:159, y = c(1, 8, 14, 27, 36, 43, 55, 69, 86, 109, 137, > >> >> 200, 245, 313, 345, 443, 558, 674, 786, 901, 989, 1075, > >> >> 1181, 1317, 1475, 1673, 1801, 1988, 2100, 2249, 2528, > >> >> 2752, 2974, 3234, 3400, 3472, 3574, 3751, 4016, 4210, > >> >> 4273, 4467, 4658, 4821, 4992, 5166, 5338, 5538, 5779, > >> >> 6021, 6200, 6378, 6532, 6720, 7090, 7197, 7387, 7523, > >> >> 7731, 7868, 8070, 8282, 8448, 8616, 8783, 8944, 9118, > >> >> 9268, 9449, 9606, 9726, 9867, 9977, 10116, 10267, 10577, > >> >> 10926, 11183, 11447, 11728, 12131, 12531, 13015, 13463, > >> >> 13837, 14095, 14609, 15044, 15463, 16004, 16425, 16854, > >> >> 17233, 17889, 18586, 19211, 20059, 20686, 21422, 21962, > >> >> 22597, 23351, 24274, 25222, 26030, 26752, 27314, 28030, > >> >> 29037, 29905, 30658, 31686, 32785, 33550, 34463, 35237, > >> >> 35995, 36983, 38149, 39334, 40291, 41251, 42216, 43257, > >> >> 44352, 45633, 47177, 48096, 49243, 50373, 51408, 52261, > >> >> 53468, 54426, 55153, 55906, 56817, 57993, 58864, 60296, > >> >> 61442, 62223, 63269, 64191, 65256, 66383, 67453, 68456, > >> >> 69424, 70231, 71418, 72560, 73651, 74492, 75394, 76464, > >> >> 77377, 78446, 79402)), row.names = c(NA, 159L), class = > >> >> "data.frame") Any help and/or guidance will be greatly > >> >> appreciated, > >> >> > >> >> Cheers, > >> >> > >> >> Paul > >> >> > >> >> [[alternative HTML version deleted]] > >> >> > >> >> ______________________________________________ > >> >> R-help@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. > >> > >> > -- > >> > Peter Dalgaard, Professor, Center for Statistics, > >> > Copenhagen Business School Solbjerg Plads 3, 2000 > >> > Frederiksberg, Denmark Phone: (+45)38153501 Office: A 4.23 > >> > Email: pd....@cbs.dk Priv: pda...@gmail.com > >> > >> > ______________________________________________ > >> > R-help@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. > >> > > > > [[alternative HTML version deleted]] > > > >______________________________________________ > >R-help@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. > > -- > Sent from my phone. Please excuse my brevity. > [[alternative HTML version deleted]] ______________________________________________ R-help@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.