eps 0.4.1 → 0.6.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4df8a83bee7fce8feebec2cf26d33d7ee4ca74fbcda9f41fb070f614cfb2e0eb
-  data.tar.gz: 23f7dd9aa63eb4306268f19b862de3a07f9d72d9ec507160a7e6d291ea2245c6
+  metadata.gz: 701cde7907172e33d54d05ee0f2236dbfa84486744e7a510190618848c878418
+  data.tar.gz: a6be730db321a5143e34727497c6ed675f8d734b6195cd4fdc5f8e6bc4e98ff1
 SHA512:
-  metadata.gz: c24ea7abf903829b3fe00dd0f7c601062464ecc193ccd8a725a98a437e7ed6f6bff8952c1c50aeeadcc5981e84325a44efedd53580e23f2475f1c8a7b927ed78
-  data.tar.gz: 601cf18d044fd9ac348d3f632b7edda7fbd34ef11f497d4be998d62ae76f33f6681953fb5c263924deebed184b5b6f560bcd24de272c509317fb9c3b68f2f3b9
+  metadata.gz: 6c76ec99a116a91cb9550b9a007b3150689b76649b14aa0c85582176fb44a08957b37c40e54ed1adb6447d8c5ec06757a611fad51a582581031f5724317ee888
+  data.tar.gz: 2176bf4351c26df4562a2879ecb2f0637d2d23bb624dc683790befa2029c2a3d89388b7d5b3183d4d48789801068d7442807c28c2a53439089e2ba0943e0bc9b

data/CHANGELOG.md

@@ -1,3 +1,11 @@
+## 0.6.0 (2025-02-01)
+
+- Dropped support for Ruby < 3.1
+
+## 0.5.0 (2023-07-02)
+
+- Dropped support for Ruby < 3
+
 ## 0.4.1 (2022-09-28)
 
 - Fixed `cannot load such file -- matrix` error with Ruby 3.1

data/LICENSE.txt

@@ -1,6 +1,6 @@
 The MIT License (MIT)
 
-Copyright (c) 2018-2021 Andrew Kane
+Copyright (c) 2018-2024 Andrew Kane
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

data/README.md

@@ -7,7 +7,7 @@ Machine learning for Ruby
 
 Check out [this post](https://ankane.org/rails-meet-data-science) for more info on machine learning with Rails
 
-[![Build Status](https://github.com/ankane/eps/workflows/build/badge.svg?branch=master)](https://github.com/ankane/eps/actions)
+[![Build Status](https://github.com/ankane/eps/actions/workflows/build.yml/badge.svg)](https://github.com/ankane/eps/actions)
 
 ## Installation
 
@@ -414,7 +414,7 @@ Eps::Model.new(data, validation_set: validation_set)
 Split on a specific value
 
 ```ruby
-Eps::Model.new(data, split: {column: :listed_at, value: Date.parse("2019-01-01")})
+Eps::Model.new(data, split: {column: :listed_at, value: Date.parse("2025-01-01")})
 ```
 
 Specify the validation set size (the default is `0.25`, which is 25%)
@@ -435,7 +435,7 @@ The database is another place you can store models. It’s good if you retrain m
 
 > We recommend adding monitoring and guardrails as well if you retrain automatically
 
-Create an ActiveRecord model to store the predictive model.
+Create an Active Record model to store the predictive model.
 
 ```sh
 rails generate model Model key:string:uniq data:text
@@ -479,61 +479,7 @@ Weights are supported for metrics as well
 Eps.metrics(actual, predicted, weight: weight)
 ```
 
-Reweighing is one method to [mitigate bias](http://aif360.mybluemix.net/) in training data
-
-## Upgrading
-
-## 0.3.0
-
-Eps 0.3.0 brings a number of improvements, including support for LightGBM and cross-validation. There are a number of breaking changes to be aware of:
-
-- LightGBM is now the default for new models. On Mac, run:
-
-  ```sh
-  brew install libomp
-  ```
-
-  Pass the `algorithm` option to use linear regression or naive Bayes.
-
-  ```ruby
-  Eps::Model.new(data, algorithm: :linear_regression) # or :naive_bayes
-  ```
-
-- Cross-validation happens automatically by default. You no longer need to create training and test sets manually. If you were splitting on a time, use:
-
-  ```ruby
-  Eps::Model.new(data, split: {column: :listed_at, value: Date.parse("2019-01-01")})
-  ```
-
-  Or randomly, use:
-
-  ```ruby
-  Eps::Model.new(data, split: {validation_size: 0.3})
-  ```
-
-  To continue splitting manually, use:
-
-  ```ruby
-  Eps::Model.new(data, validation_set: test_set)
-  ```
-
-- It’s no longer possible to load models in JSON or PFA formats. Retrain models and save them as PMML.
-
-## 0.2.0
-
-Eps 0.2.0 brings a number of improvements, including support for classification.
-
-We recommend:
-
-1. Changing `Eps::Regressor` to `Eps::Model`
-2. Converting models from JSON to PMML
-
-  ```ruby
-  model = Eps::Model.load_json("model.json")
-  File.write("model.pmml", model.to_pmml)
-  ```
-
-3. Renaming `app/stats_models` to `app/ml_models`
+Reweighing is one method to [mitigate bias](https://fairlearn.org/) in training data
 
 ## History
 

data/lib/eps/base_estimator.rb

@@ -226,7 +226,7 @@ module Eps
         end
 
         encoder.vocabulary.each do |word|
-          train_set.columns[[k, word]] = [0] * counts.size
+          train_set.columns[[k, word]] = Array.new(counts.size, 0)
         end
 
         counts.each_with_index do |ci, i|

data/lib/eps/data_frame.rb

@@ -54,7 +54,7 @@ module Eps
     def map
       if @columns.any?
         size.times.map do |i|
-          yield Hash[@columns.map { |k, v| [k, v[i]] }]
+          yield @columns.to_h { |k, v| [k, v[i]] }
         end
       end
     end

data/lib/eps/evaluators/linear_regression.rb

@@ -4,7 +4,7 @@ module Eps
       attr_reader :features
 
       def initialize(coefficients:, features:, text_features:)
-        @coefficients = Hash[coefficients.map { |k, v| [k.is_a?(Array) ? [k[0].to_s, k[1]] : k.to_s, v] }]
+        @coefficients = coefficients.to_h { |k, v| [k.is_a?(Array) ? [k[0].to_s, k[1]] : k.to_s, v] }
         @features = features
         @text_features = text_features || {}
       end
@@ -13,7 +13,7 @@ module Eps
         raise "Probabilities not supported" if probabilities
 
         intercept = @coefficients["_intercept"] || 0.0
-        scores = [intercept] * x.size
+        scores = Array.new(x.size, intercept)
 
         @features.each do |k, type|
           raise "Missing data in #{k}" if !x.columns[k] || x.columns[k].any?(&:nil?)
@@ -50,7 +50,7 @@ module Eps
       end
 
       def coefficients
-        Hash[@coefficients.map { |k, v| [Array(k).join.to_sym, v] }]
+        @coefficients.to_h { |k, v| [Array(k).join.to_sym, v] }
       end
     end
   end

data/lib/eps/evaluators/naive_bayes.rb

@@ -33,7 +33,7 @@ module Eps
         total = probabilities[:prior].values.sum.to_f
         probabilities[:prior].each do |c, cv|
           prior = Math.log(cv / total)
-          px = [prior] * x.size
+          px = Array.new(x.size, prior)
 
           @features.each do |k, type|
             case type

data/lib/eps/label_encoder.rb

@@ -36,7 +36,7 @@ module Eps
     end
 
     def inverse_transform(y)
-      inverse = Hash[@labels.map(&:reverse)]
+      inverse = @labels.map(&:reverse).to_h
       y.map do |yi|
         inverse[yi.to_i]
       end

data/lib/eps/linear_regression.rb

@@ -146,7 +146,7 @@ module Eps
 
       @coefficient_names = data.columns.keys
       @coefficient_names.unshift("_intercept") if intercept
-      @coefficients = Hash[@coefficient_names.zip(v3)]
+      @coefficients = @coefficient_names.zip(v3).to_h
       Evaluators::LinearRegression.new(coefficients: @coefficients, features: @features, text_features: @text_features)
     end
 
@@ -172,21 +172,20 @@ module Eps
     # add epsilon for perfect fits
     # consistent with GSL
     def t_value
-      @t_value ||= Hash[@coefficients.map { |k, v| [k, v / (std_err[k] + Float::EPSILON)] }]
+      @t_value ||= @coefficients.to_h { |k, v| [k, v / (std_err[k] + Float::EPSILON)] }
     end
 
     def p_value
       @p_value ||= begin
-        Hash[@coefficients.map do |k, _|
-          tp = Eps::Statistics.tdist_p(t_value[k].abs, degrees_of_freedom)
-          [k, 2 * (1 - tp)]
-        end]
+        @coefficients.to_h do |k, _|
+          [k, 2 * Eps::Statistics.students_t_cdf(-t_value[k].abs, degrees_of_freedom)]
+        end
       end
     end
 
     def std_err
       @std_err ||= begin
-        Hash[@coefficient_names.zip(diagonal.map { |v| Math.sqrt(v) })]
+        @coefficient_names.zip(diagonal.map { |v| Math.sqrt(v) }).to_h
       end
     end
 

data/lib/eps/statistics.rb

@@ -1,79 +1,73 @@
-### Extracted from https://github.com/estebanz01/ruby-statistics
-### The Ruby author is Esteban Zapata Rojas
-###
-### Originally extracted from https://codeplea.com/incomplete-beta-function-c
-### These functions shared under zlib license and the author is Lewis Van Winkle
-
 module Eps
   module Statistics
-    def self.tdist_p(value, degrees_of_freedom)
-      upper = (value + Math.sqrt(value * value + degrees_of_freedom))
-      lower = (2.0 * Math.sqrt(value * value + degrees_of_freedom))
-
-      x = upper/lower
-
-      alpha = degrees_of_freedom/2.0
-      beta = degrees_of_freedom/2.0
-
-      incomplete_beta_function(x, alpha, beta)
+    def self.normal_cdf(x, mean, std_dev)
+      0.5 * (1.0 + Math.erf((x - mean) / (std_dev * Math.sqrt(2))))
     end
 
-    def self.incomplete_beta_function(x, alp, bet)
-      return if x < 0.0
-      return 1.0 if x > 1.0
-
-      tiny = 1.0E-50
-
-      if x > ((alp + 1.0)/(alp + bet + 2.0))
-        return 1.0 - incomplete_beta_function(1.0 - x, bet, alp)
+    # Hill, G. W. (1970).
+    # Algorithm 395: Student's t-distribution.
+    # Communications of the ACM, 13(10), 617-619.
+    def self.students_t_cdf(x, n)
+      start, sign = x < 0 ? [0, 1] : [1, -1]
+
+      z = 1.0
+      t = x * x
+      y = t / n.to_f
+      b = 1.0 + y
+
+      if n > n.floor || (n >= 20.0 && t < n) || n > 200.0
+        # asymptotic series for large or noninteger n
+        if y > 10e-6
+          y = Math.log(b)
+        end
+        a = n - 0.5
+        b = 48.0 * a * a
+        y *= a
+        y = (((((-0.4 * y - 3.3) * y - 24.0) * y - 85.5) / (0.8 * y * y + 100.0 + b) + y + 3.0) / b + 1.0) * Math.sqrt(y)
+        return start + sign * normal_cdf(-y, 0.0, 1.0)
       end
 
-      # To avoid overflow problems, the implementation applies the logarithm properties
-      # to calculate in a faster and safer way the values.
-      lbet_ab = (Math.lgamma(alp)[0] + Math.lgamma(bet)[0] - Math.lgamma(alp + bet)[0]).freeze
-      front = (Math.exp(Math.log(x) * alp + Math.log(1.0 - x) * bet - lbet_ab) / alp.to_f).freeze
-
-      # This is the non-log version of the left part of the formula (before the continuous fraction)
-      # down_left = alp * self.beta_function(alp, bet)
-      # upper_left = (x ** alp) * ((1.0 - x) ** bet)
-      # front = upper_left/down_left
-
-      f, c, d = 1.0, 1.0, 0.0
-
-      returned_value = nil
-
-      # Let's do more iterations than the proposed implementation (200 iters)
-      (0..500).each do |number|
-        m = number/2
-
-        numerator = if number == 0
-                      1.0
-                    elsif number % 2 == 0
-                      (m * (bet - m) * x)/((alp + 2.0 * m - 1.0)* (alp + 2.0 * m))
-                    else
-                      top = -((alp + m) * (alp + bet + m) * x)
-                      down = ((alp + 2.0 * m) * (alp + 2.0 * m + 1.0))
-
-                      top/down
-                    end
-
-        d = 1.0 + numerator * d
-        d = tiny if d.abs < tiny
-        d = 1.0 / d
-
-        c = 1.0 + numerator / c
-        c = tiny if c.abs < tiny
-
-        cd = (c*d).freeze
-        f = f * cd
+      if n < 20 && t < 4.0
+        # nested summation of cosine series
+        y = Math.sqrt(y)
+        a = y
+        if n == 1
+          a = 0.0
+        end
 
-        if (1.0 - cd).abs < 1.0E-10
-          returned_value = front * (f - 1.0)
-          break
+        # loop
+        if n > 1
+          n -= 2
+          while n > 1
+            a = (n - 1) / (b * n) * a + y
+            n -= 2
+          end
         end
+        a = n == 0 ? a / Math.sqrt(b) : (Math.atan(y) + a / b) * (2.0 / Math::PI)
+        return start + sign * (z - a) / 2.0
       end
 
-      returned_value
+      # tail series expanation for large t-values
+      a = Math.sqrt(b)
+      y = a * n
+      j = 0
+      while a != z
+        j += 2
+        z = a
+        y = y * (j - 1) / (b * j)
+        a += y / (n + j)
+      end
+      z = 0.0
+      y = 0.0
+      a = -a
+
+      # loop (without n + 2 and n - 2)
+      while n > 1
+        a = (n - 1) / (b * n) * a + y
+        n -= 2
+      end
+      a = n == 0 ? a / Math.sqrt(b) : (Math.atan(y) + a / b) * (2.0 / Math::PI)
+      start + sign * (z - a) / 2.0
     end
   end
 end

data/lib/eps/text_encoder.rb

@@ -27,7 +27,7 @@ module Eps
 
       max_features = options[:max_features]
       if max_features
-        counts = Hash[counts.sort_by { |_, v| -v }[0...max_features]]
+        counts = counts.sort_by { |_, v| -v }[0...max_features].to_h
       end
 
       @vocabulary = counts.keys

data/lib/eps/version.rb

@@ -1,3 +1,3 @@
 module Eps
-  VERSION = "0.4.1"
+  VERSION = "0.6.0"
 end

data/lib/eps.rb

@@ -1,34 +1,36 @@
 # dependencies
-require "json"
 require "lightgbm"
 require "matrix"
 require "nokogiri"
 
+# stdlib
+require "json"
+
 # modules
-require "eps/base"
-require "eps/base_estimator"
-require "eps/data_frame"
-require "eps/label_encoder"
-require "eps/lightgbm"
-require "eps/linear_regression"
-require "eps/metrics"
-require "eps/model"
-require "eps/naive_bayes"
-require "eps/statistics"
-require "eps/text_encoder"
-require "eps/utils"
-require "eps/version"
+require_relative "eps/base"
+require_relative "eps/base_estimator"
+require_relative "eps/data_frame"
+require_relative "eps/label_encoder"
+require_relative "eps/lightgbm"
+require_relative "eps/linear_regression"
+require_relative "eps/metrics"
+require_relative "eps/model"
+require_relative "eps/naive_bayes"
+require_relative "eps/statistics"
+require_relative "eps/text_encoder"
+require_relative "eps/utils"
+require_relative "eps/version"
 
 # pmml
-require "eps/pmml"
-require "eps/pmml/generator"
-require "eps/pmml/loader"
+require_relative "eps/pmml"
+require_relative "eps/pmml/generator"
+require_relative "eps/pmml/loader"
 
 # evaluators
-require "eps/evaluators/linear_regression"
-require "eps/evaluators/lightgbm"
-require "eps/evaluators/naive_bayes"
-require "eps/evaluators/node"
+require_relative "eps/evaluators/linear_regression"
+require_relative "eps/evaluators/lightgbm"
+require_relative "eps/evaluators/naive_bayes"
+require_relative "eps/evaluators/node"
 
 module Eps
   class Error < StandardError; end

metadata

@@ -1,14 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: eps
 version: !ruby/object:Gem::Version
-  version: 0.4.1
+  version: 0.6.0
 platform: ruby
 authors:
 - Andrew Kane
-autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-09-28 00:00:00.000000000 Z
+date: 2025-02-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: lightgbm
@@ -52,7 +51,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-description:
 email: andrew@ankane.org
 executables: []
 extensions: []
@@ -86,7 +84,6 @@ homepage: https://github.com/ankane/eps
 licenses:
 - MIT
 metadata: {}
-post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -94,15 +91,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '2.7'
+      version: '3.1'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.7
-signing_key:
+rubygems_version: 3.6.2
 specification_version: 4
 summary: Machine learning for Ruby. Supports regression (linear regression) and classification
   (naive Bayes)