dkam/decisiontree
1
0
Fork 0
mirror of https://github.com/dkam/decisiontree.git synced 2025-12-27 22:54:52 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Simplify with #sum

Browse Source
This commit is contained in:
Brian Underwood
2017-04-11 14:57:32 -04:00
parent 868ed91672
commit 13aed0b2ae
3 changed files with 5 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

BIN
lib/.DS_Store vendored Normal file
View File

Binary file not shown.

5
lib/core_extensions/array.rb
View File

@@ -2,9 +2,10 @@ class Array
def entropy def entropy
each_with_object(Hash.new(0)) do |i, result| each_with_object(Hash.new(0)) do |i, result|
result[i] += 1 result[i] += 1
end.values.inject(0) do |sum, count| end.values.sum do |count|
percentage = count.to_f / length percentage = count.to_f / length
sum + -percentage * Math.log2(percentage)
-percentage * Math.log2(percentage)
end end
end end
end end

4
lib/decisiontree/id3_tree.rb
View File

@@ -120,12 +120,12 @@ module DecisionTree
index = attributes.index(attribute) index = attributes.index(attribute)
values = data.map { |row| row[index] }.uniq values = data.map { |row| row[index] }.uniq
remainder = values.sort.inject(0) do |sum, val| remainder = values.sort.sum do |val|
classification = data.each_with_object([]) do |row, result| classification = data.each_with_object([]) do |row, result|
result << row.last if row[index] == val result << row.last if row[index] == val
end end
sum + ((classification.size.to_f / data.size) * classification.entropy) ((classification.size.to_f / data.size) * classification.entropy)
end end
[data.classification.entropy - remainder, index] [data.classification.entropy - remainder, index]