Move interface definitions into packing modules.

Makefile

@@ -4,7 +4,7 @@ DOC_BUILD_DIR="_doc_build"
 PACKAGES_INSTALL=ocephes lacaml lbfgs
 PACKAGES_INSTALL_TEST=kaputt bisect_ppx dsfo
 
-.PHONY: all clean test build full setup doc
+.PHONY: all clean test build full setup doc examples examples_lite
 
 default: FORCE
 	@echo "available targets:"
@@ -19,6 +19,7 @@ default: FORCE
 	@echo "	doc			generates ocamldoc documentations"
 	@echo "	clean			deletes all produced files"
 	@echo "	report			generate Bisect_ppx coverage report"
+	@echo " showed  generate showed.mli that contains the current interfaces"
 
 # This should be called something else.
 setup:
@@ -36,7 +37,6 @@ build:
 build_lite:
 	ocaml pkg/pkg.ml build --with-lacaml false --with-lbfgs false --with-ocephes false
 
-
 clean:
 	ocaml pkg/pkg.ml clean
 	ocaml pkg/pkg.ml clean --build-dir $(TEST_BUILD_DIR)
@@ -61,6 +61,7 @@ covered_test_lite:
 	ocaml pkg/pkg.ml build --build-dir $(COVERED_TEST_BUILD_DIR) --with-lacaml false --with-lbfgs false --with-ocephes false --with-coverage true -n omltest && \
 		time ocaml pkg/pkg.ml test --build-dir $(COVERED_TEST_BUILD_DIR)
 
+
 #### Test Coverage
 
 report_dir:
@@ -82,5 +83,19 @@ clean_reports:
 doc:
 	ocamlbuild -classic-display -use-ocamlfind -plugin-tag 'package(str)' -no-links -build-dir $(DOC_BUILD_DIR) -docflags '-colorize-code,-charset,utf-8' doc/api.docdir/index.html
 
-# topkg doc --build-dir $(DOC_BUILD_DIR)
+showed: build
+	utop -I _build/src oml.cma tools/show.ml > showed.mli
+	utop -require lacaml -require lbfgs -require ocephes -I _build/src oml.cma -I _build/src-full oml_full.cma tools/show_full.ml > showed_full.mli
+
+#### Examples
+
+examples_lite:
+	ocaml pkg/pkg.ml build --build-dir $(TEST_BUILD_DIR) --with-lacaml false --with-lbfgs false --with-ocephes false -n examples && \
+		time ocaml pkg/pkg.ml test --build-dir $(TEST_BUILD_DIR)
+
+examples:
+	ocaml pkg/pkg.ml build --build-dir $(TEST_BUILD_DIR) -n examples && \
+		time ocaml pkg/pkg.ml test --build-dir $(TEST_BUILD_DIR)
+
 FORCE:
+

_tags

@@ -17,6 +17,8 @@ true: safe_string, annot, bin_annot, principal, warn(A-42-44-48-50)
 <src-full/**/*.mlj> : package(kaputt), package(dsfo)
 <test/**> : package(dsfo), package(kaputt)
 <test/omlf_test.native> : package(lacaml), package(ocephes), package(lbfgs)
+<examples/*> : include
+<examples/oml_full/*.native> : package(lacaml), package(ocephes), package(lbfgs)
 
 "_build": -traverse
 "_test_build": -traverse

scripts/examples/classify_example.ml → examples/oml/classify.ml

@@ -1,4 +1,7 @@
+open Oml
+open Util
 open Classification
+
 let feature_map = function | `shortbread -> 0 | `lager -> 1 | `whiskey -> 2 | `porridge -> 3 | `football-> 4 ;;
 let data =
   [
@@ -20,10 +23,15 @@ let to_feature l = l |> List.map ~f:feature_map |> Array.of_list ;;
 module NB = Naive_bayes.Binomial(
   struct
     type feature = [ `shortbread | `lager | `whiskey | `porridge | `football ] list
-    type clas = [ `English | `Scottish]
+    type class_ = [ `English | `Scottish]
     let encoding = to_feature
     let size = 5
   end);;
 let naiveb = NB.estimate ~opt:( NB.opt ~bernoulli:true ()) data ;;
 let sample = [ `shortbread ; `whiskey; `porridge  ]  ;;
 let result = NB.eval naiveb sample;;
+
+let () =
+  if !Sys.interactive then () else
+    Printf.printf "Probability of English: %f\t(expecting 19%%)\n" (List.assoc `English result);
+    Printf.printf "Probability of Scottish: %f\t(expecting 81%%)\n" (List.assoc `Scottish result)

examples/oml/generate_fat_tailed_data.ml

@@ -0,0 +1,18 @@
+open Oml.Util
+open Oml.Statistics
+
+let n1 = Sampling.normal ~mean:2. ~std:1. ()
+let n2 = Sampling.normal ~mean:2. ~std:10. ()
+let samples = 10000
+let data = Array.init samples (fun i -> if i mod 10 = 0 then n2 () else n1 ())
+let kurtosis = Descriptive.classify_kurtosis data
+
+let () =
+  if !Sys.interactive then () else
+    Printf.printf "Our data has %s kurtosis\t (expecting fat).\n"
+      (match kurtosis with
+        | `Fat -> "fat"
+        | `Normal -> "normal"
+        | `Skinny -> "skinny"
+        | `Slightly_fat -> "slightly fat"
+        | `Slightly_skinny -> "slightly skinny")

examples/oml/skewness.ml

@@ -0,0 +1,28 @@
+open Oml
+open Statistics
+
+(* TODO: Redo this example with a distribution that better models skew. *)
+let samples = 10000
+let n1 = Sampling.normal ~mean:2. ~std:1. ()
+let n2 = Sampling.normal ~mean:3. ~std:1. ()
+
+let data_skew ftf =
+  Array.init samples (fun i -> if i mod ftf = 0 then n2 () else n1())
+  |> Descriptive.summary
+  |> fun s -> s.Descriptive.skew
+  |> snd ;;
+
+let count_big_skew =
+  Array.fold_left (fun c s ->
+      if s = `Positive || s = `Negative then c + 1 else c) 0 ;;
+
+let test_size = 100 ;;
+
+let t10 = Array.init test_size (fun _ -> data_skew 10) |> count_big_skew ;;
+let t2 = Array.init test_size (fun _ -> data_skew 2) |> count_big_skew ;;
+
+let () =
+  if !Sys.interactive then () else
+    Printf.printf "We observed skewed data %d out of %d when the frequency of drawing from a larger mean was 1/%d.\n\
+                   But only %d out of %d when the frequency was 1/%d.\n"
+      t10 test_size 10 t2 test_size 2

examples/oml_full/mean_test.ml

@@ -0,0 +1,18 @@
+open Oml_full
+open Statistics
+
+let mean = 2.
+let std = 1.
+let gen = Sampling.normal ~mean ~std () ;;
+let samples = 100
+
+let data = Array.init samples (fun _ -> gen ()) ;;
+let test_value = 1.9
+let tr = Hypothesis_test.mean_t_test test_value Hypothesis_test.Two_sided data ;;
+
+let () =
+  if !Sys.interactive then () else
+    Printf.printf "Our data is %d samples with a mean of %f and a std of %f, \
+          if we test for a mean of %f \n\
+          the prob of seeing this result by chance is: %f\n"
+      samples mean std test_value tr.Hypothesis_test.prob_by_chance

examples/oml_full/regression_ex.ml

@@ -0,0 +1,30 @@
+open Oml_full
+open Statistics
+open Regression
+
+let rows = 100
+let cols = 3
+
+let gen = Sampling.normal_std () ;;
+let pred = Array.init rows (fun _ -> Array.init cols (fun _ -> gen ())) ;;
+let constant_term = 2.
+let a1 = 3.
+let a2 = 4.
+let a3 = 5.
+let resp =
+  Array.map (fun row ->
+    constant_term +. a1 *. row.(0) +. a2 *. row.(1) +. a3 *. row.(2)) pred ;;
+
+let opt = (* Redundant, just for examples *)
+  Multivariate.opt ~add_constant_column:true
+    () ;;
+
+let t = Multivariate.regress ~opt pred ~resp ;;
+let c = Multivariate.coefficients t ;;
+
+let () =
+  if !Sys.interactive then () else
+    Printf.printf "Regression a linear model of y = %f + %f * x1 + %f * x2 + %f * x3 \n\
+      over randomly generated input data, our regression coefficients are: \n\
+        %f %f %f %f."
+        constant_term a1 a2 a3 c.(0) c.(1) c.(2) c.(3)

myocamlbuild.ml

@@ -219,7 +219,7 @@ let () =
       | After_rules     ->
           begin
             let regular_source_dirs =
-                  [ "src"         (* For the Online stuff that hasn't been packaged. *)
+                  [ "src/onl"
                   ; "src/util"
                   ; "src/unc"
                   ; "src/stats"
@@ -229,8 +229,7 @@ let () =
                   ]
             in
             let full_source_dirs =
-                  [ "src-full"
-                  ; "src-full/unc"
+                  [ "src-full/unc"
                   ; "src-full/stats"
                   ; "src-full/cls"
                   ; "src-full/rgr"

pkg/pkg.ml

@@ -47,5 +47,12 @@ let () =
         Ok  [ Pkg.test ~cond:(not full) "test/oml_test"
             ; Pkg.test ~cond:full "test/omlf_test"
             ]
+    | "examples" ->
+        Ok  [ Pkg.test  "examples/oml/classify"
+            ; Pkg.test  "examples/oml/generate_fat_tailed_data"
+            ; Pkg.test  "examples/oml/skewness"
+            ; Pkg.test ~cond:full "examples/oml_full/mean_test"
+            ; Pkg.test ~cond:full "examples/oml_full/regression_ex"
+            ]
     | other ->
         R.error_msgf "Unrecognized package name: %s" other

scripts/examples/float_format.ml → scripts/float_format.ml

@@ -1,2 +1,4 @@
+(* When you want print shorter floats *)
+
 let shorter_float_printer fr = Format.fprintf fr "%0.4f"  ;;
 #install_printer shorter_float_printer ;;

scripts/how_good.ml

@@ -0,0 +1,108 @@
+(** Algorithm 241 
+   http://lib.stat.cmu.edu/apstat/241
+   *)
+
+module Float = struct
+  let ( + ) x y = x +. y
+  let ( - ) x y = x -. y
+  let ( * ) x y = x *. y
+  let ( / ) x y = x /. y
+end
+
+(**
+DOUBLE PRECISION FUNCTION PPND16 (P, IFAULT)
+
+	ALGORITHM AS241  APPL. STATIST. (1988) VOL. 37, NO. 3
+
+	Produces the normal deviate Z corresponding to a given lower
+	tail area of P; Z is accurate to about 1 part in 10**16.
+*)
+let ppnd16 p =
+  (* Coefficients for P close to 0.5 *)
+  let a =
+    [| 3.3871328727963666080e+0
+     ; 1.3314166789178437745e+2
+     ; 1.9715909503065514427e+3
+     ; 1.3731693765509461125e+4
+     ; 4.5921953931549871457e+4
+     ; 6.7265770927008700853e+4
+     ; 3.3430575583588128105e+4
+     ; 2.5090809287301226727e+3
+    |] in
+let b =
+    [| 0.
+     ; 4.2313330701600911252e+1
+     ; 6.8718700749205790830e+2
+     ; 5.3941960214247511077e+3
+     ; 2.1213794301586595867e+4
+     ; 3.9307895800092710610e+4
+     ; 2.8729085735721942674e+4
+     ; 5.2264952788528545610e+3
+    |] in
+  let c =
+    [| 1.42343711074968357734e0
+     ; 4.63033784615654529590e0
+     ; 5.76949722146069140550e0
+     ; 3.64784832476320460504e0
+     ; 1.27045825245236838258e0
+     ; 2.41780725177450611770e-1
+     ; 2.27238449892691845833e-2
+     ; 7.74545014278341407640e-4
+     |] in
+  let d =
+    [| 0.
+     ; 2.05319162663775882187e0
+     ; 1.67638483018380384940e0
+     ; 6.89767334985100004550e-1
+     ; 1.48103976427480074590e-1
+     ; 1.51986665636164571966e-2
+     ; 5.47593808499534494600e-4
+     ; 1.05075007164441684324e-9
+     |] in
+  let e =
+    [| 6.65790464350110377720e0
+     ; 5.46378491116411436990e0
+     ; 1.78482653991729133580e0
+     ; 2.96560571828504891230e-1
+     ; 2.65321895265761230930e-2
+     ; 1.24266094738807843860e-3
+     ; 2.71155556874348757815e-5
+     ; 2.01033439929228813265e-7
+    |] in
+  let f =
+    [| 0.
+     ; 5.99832206555887937690e-1
+     ; 1.36929880922735805310e-1
+     ; 1.48753612908506148525e-2
+     ; 7.86869131145613259100e-4
+     ; 1.84631831751005468180e-5
+     ; 1.42151175831644588870e-7
+     ; 2.04426310338993978564e-15
+    |] in
+  let open Float in
+  let q = p - 0.5 in
+  if abs_float q < 0.425 then
+    let r = 0.180625 - q * q in
+    q * (((((((a.(7) * r + a.(6)) * r + a.(5)) * r + a.(4)) * r + a.(3))
+     			      * r + a.(2)) * r + a.(1)) * r + a.(0)) /
+     		        (((((((b.(7) * r + b.(6)) * r + b.(5)) * r + b.(4)) * r + b.(3))
+     			      * r + b.(2)) * r + b.(1)) * r + 1.)
+  else
+    let inv p = if q < 0. then -1. * p else p in
+    let r = if q < 0.0 then p else 1. - p in
+    if r < 0. then 0. else
+      let r = sqrt (-1. * log r) in
+      if r < 5.0 then
+        let r = r - 1.6 in
+        (((((((c.(7) * r + c.(6)) * r + c.(5)) * r + c.(4)) * r + c.(3))
+     			      * r + c.(2)) * r + c.(1)) * r + c.(0)) /
+     		       (((((((d.(7) * r + d.(6)) * r + d.(5)) * r + d.(4)) * r + d.(3))
+     			      * r + d.(2)) * r + d.(1)) * r + 1.)
+        |> inv
+      else
+        let r = r - 5.0 in
+	      (((((((e.(7) * r + e.(6)) * r + e.(5)) * r + e.(3)) * r + e.(3))
+     			      * r + e.(2)) * r + e.(1)) * r + e.(0)) /
+     		      (((((((f.(7) * r + f.(6)) * r + f.(5)) * r + f.(4)) * r + f.(3))
+     			      * r + f.(2)) * r + f.(1)) * r + 1.)
+        |> inv