Problems with linear model in Incanter 1.9.2 about incanter

[joinr]

Note:

I got the vector example (not using clatrix), 0 problems on my joinr/incanter 1.9.3-SNAPSHOT fork, and the deployed incanter 1.9.3 on clojars (a bit behind mine, but compatible):

user> (use 'incanter.core)
nil
user> (require 'incanter.stats)
nil
user> (incanter.stats/linear-model [1 2 3] [1 2 3])
{:y [1 2 3], :sse 9.466330862652142E-30, :msr 2.0, :design-matrix #vectorz/array [[1.0,1.0],[1.0,2.0],[1.0,3.0]], :mse 9.466330862652142E-30, :t-probs [0.769946543837384 1.3322676295501878E-15], :adj-r-square NaN, :df [1 1], :coef-var #vectorz/matrix [[2.2088105346188328E-29,-9.466330862652142E-30],
[-9.466330862652142E-30,4.733165431326071E-30]], :residuals (1.7763568394002505E-15 1.7763568394002505E-15 1.7763568394002505E-15), :ssr 2.0, :sst 2.0, :coefs (-1.7763568394002505E-15 1.0), :f-stat 2.1127510003803822E29, :r-square 1.0, :f-prob 0.0, :t-tests [-0.37796447300922725 4.5964671220192394E14], :x #vectorz/array [[1.0,1.0],[1.0,2.0],[1.0,3.0]], :std-errors #vectorz/vector [4.6997984367617645E-15,2.175583928816829E-15], :fitted (0.9999999999999982 1.9999999999999982 2.9999999999999982), :coefs-ci ((-6.14929580020137E-14 5.79402443232132E-14) (0.9999999999999724 1.0000000000000275))}

[joinr]

update:

you need to include clatrix 0.5.0 in your dependencies, with the following caveat (based off of what incanter does):

[clatrix "0.5.0" :exclusions [org.clojure/clojure net.mikera/core.matrix]]

That version works with 1.9.3 and your example. It basically says (don't use the clojure and core matrix deps that clatrix bundles, since they're old 1.7alpha and 0.36.1, incanter is using 1.8 and 0.52 respectively). Clatrix is a dev dependency for incanter.core, so it's not on the classpath by default.

[joinr]

A general approach that deviates from the example is to use the built-in incanter.core/matrix function.
This will delegate to the bundled vectorz library and works for n-dimensional vectors.

(def X (incanter.core/matrix [8.401 14.475 13.396 12.127 5.044
8.339 15.692 17.108 9.253 12.029]))
(def Y (incanter.core/matrix [-1.57 2.32 0.424 0.814 -2.3 0.01 1.954 2.296 -0.635 0.328]))

user> (incanter.stats/linear-model Y X )
{:y #vectorz/vector [-1.57,2.32,0.424,0.814,-2.3,0.01,1.954,2.296,-0.635,0.328], :sse 2.586225034528488, :msr 19.66839986547147, :design-matrix #vectorz/array [[1.0,8.401],[1.0,14.475],[1.0,13.396],[1.0,12.127],[1.0,5.044],[1.0,8.339],[1.0,15.692],[1.0,17.108],[1.0,9.253],[1.0,12.029]], :mse 0.323278129316061, :t-probs [1.3052188908235784E-4 5.2372232961728216E-5], :adj-r-square 0.8505862719364365, :df [1 8], :coef-var #vectorz/matrix [[0.3703451108456678,-0.029173625795248025],
[-0.029173625795248025,0.00251791978485535]], :residuals (-0.6873445559690559 0.8253111334125083 -0.6483716931997257 0.2383108767994172 -0.10342541689330798 0.9169220471357094 -0.017018801724573818 -0.22923670489146586 -0.08581465024744017 -0.20933223442208276), :ssr 19.66839986547147, :sst 22.25462489999996, :coefs (-4.170780164726611 0.3913968242704043), :f-stat 60.840490221477886, :r-square 0.8837893226172284, :f-prob 5.237223296161719E-5, :t-tests [-6.853524847485774 7.800031424390413], :x #vectorz/array [[1.0,8.401],[1.0,14.475],[1.0,13.396],[1.0,12.127],[1.0,5.044],[1.0,8.339],[1.0,15.692],[1.0,17.108],[1.0,9.253],[1.0,12.029]], :std-errors #vectorz/vector [0.6085598662791261,0.05017887787560967], :fitted (-0.8826554440309442 1.4946888665874916 1.0723716931997256 0.5756891232005827 -2.196574583106692 -0.9069220471357093 1.9710188017245738 2.5252367048914657 -0.5491853497525598 0.5373322344220828), :coefs-ci ((-5.574121732978319 -2.7674385964749026) (0.2756841243816959 0.5071095241591126))}