multi-armed-bandit 0.0.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/.gitignore +17 -0
- data/Gemfile +3 -0
- data/LICENSE.txt +22 -0
- data/README.md +32 -0
- data/Rakefile +1 -0
- data/lib/multi-armed-bandit.rb +7 -0
- data/lib/multi-armed-bandit/thompson_sampling.rb +126 -0
- data/lib/multi-armed-bandit/version.rb +3 -0
- data/multi-armed-bandit.gemspec +20 -0
- metadata +66 -0
checksums.yaml
ADDED
@@ -0,0 +1,7 @@
|
|
1
|
+
---
|
2
|
+
SHA1:
|
3
|
+
metadata.gz: 366b2c8ed586d54d3710a4124f3e0d27598ee162
|
4
|
+
data.tar.gz: 21592d2c199e013b35e0def04742b9e857aa0de7
|
5
|
+
SHA512:
|
6
|
+
metadata.gz: cdbabf315694cc7a90ffc522a45ce060cffbff28cbefa90cca151b666c7685ea0d12dab07d8827d4e4b6f3576084430b2df1d64f4a7b74f917bd2d478f0b8e0a
|
7
|
+
data.tar.gz: 515b9877b81c8a7c85d154baa8bab15857458a12f0745d5df01c8d4f3fe191f0e114cbf5dc71f40d20a72879cf60f9259378d82646d30f1ebe6b6e4fe08e4db7
|
data/.gitignore
ADDED
data/Gemfile
ADDED
data/LICENSE.txt
ADDED
@@ -0,0 +1,22 @@
|
|
1
|
+
Copyright (c) 2013 David Dai
|
2
|
+
|
3
|
+
MIT License
|
4
|
+
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining
|
6
|
+
a copy of this software and associated documentation files (the
|
7
|
+
"Software"), to deal in the Software without restriction, including
|
8
|
+
without limitation the rights to use, copy, modify, merge, publish,
|
9
|
+
distribute, sublicense, and/or sell copies of the Software, and to
|
10
|
+
permit persons to whom the Software is furnished to do so, subject to
|
11
|
+
the following conditions:
|
12
|
+
|
13
|
+
The above copyright notice and this permission notice shall be
|
14
|
+
included in all copies or substantial portions of the Software.
|
15
|
+
|
16
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
|
17
|
+
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
18
|
+
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
|
19
|
+
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
|
20
|
+
LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
|
21
|
+
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
|
22
|
+
WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
data/README.md
ADDED
@@ -0,0 +1,32 @@
|
|
1
|
+
# MultiArmedBandit
|
2
|
+
A Redis backed multi-armed bandit library. Currently Thompson Sampling is implemented.
|
3
|
+
|
4
|
+
## Usage
|
5
|
+
|
6
|
+
```ruby
|
7
|
+
require 'multi-armed-bandit'
|
8
|
+
mab = MultiArmedBandit::ThompsonSampling.new Redis.new(:port => 6379), 'colors'
|
9
|
+
mab.create! ['red', 'green', 'blue'], :alpha => 10, :beta => 10
|
10
|
+
mab.draw # => 'red'
|
11
|
+
mab.draw # => 'green'
|
12
|
+
mab.draw # => 'blue'
|
13
|
+
mab.draw_multi(3) # => ['red', blue', 'green']
|
14
|
+
|
15
|
+
mab.put 'yellow', :alpha => 5, :beta => 5
|
16
|
+
mab.draw # => 'yellow'
|
17
|
+
mab.update_success('red')
|
18
|
+
mab.update_success('green', 2)
|
19
|
+
mab.draw # => 'green'
|
20
|
+
mab.remove 'green'
|
21
|
+
mab.draw # => 'red'
|
22
|
+
mab.disable 'red'
|
23
|
+
mab.draw # => 'yellow'
|
24
|
+
mab.enable 'red'
|
25
|
+
mab.draw # => 'red'
|
26
|
+
```
|
27
|
+
|
28
|
+
```ruby
|
29
|
+
mab = MultiArmedBandit::ThompsonSampling.new Redis.new(:port => 6379), 'colors'
|
30
|
+
mab.load!
|
31
|
+
mab.stats # => {:arms=>["yellow", "red", "blue"], :state=>{"alpha"=>"5", "beta"=>"5", "red:count"=>"3", "red:success"=>"1", "blue:count"=>"2", "blue:success"=>"0.0", "yellow:count"=>"1", "yellow:success"=>"0.0", "green:count"=>"1"}, :means=>[["blue", 0.45454545454545453], ["green", 0.47619047619047616], ["yellow", 0.47619047619047616], ["red", 0.4782608695652174]]}
|
32
|
+
```
|
data/Rakefile
ADDED
@@ -0,0 +1 @@
|
|
1
|
+
require "bundler/gem_tasks"
|
@@ -0,0 +1,126 @@
|
|
1
|
+
module MultiArmedBandit
|
2
|
+
class ThompsonSampling
|
3
|
+
def self.beta_mean(success, count, alpha, beta)
|
4
|
+
fail_count = count - success
|
5
|
+
1 / (1 + (fail_count+beta).to_f / (success+alpha))
|
6
|
+
end
|
7
|
+
|
8
|
+
attr_reader :name, :arms
|
9
|
+
attr_accessor :redis
|
10
|
+
|
11
|
+
def initialize(redis, name, options={})
|
12
|
+
@redis, @name, = redis, name
|
13
|
+
@arms_key, @means_key = "#{@name}:arms", "#{@name}:means"
|
14
|
+
end
|
15
|
+
|
16
|
+
def create!(arms, options={})
|
17
|
+
@arms = Set.new arms
|
18
|
+
ts = {:alpha=>5, :beta=>5}
|
19
|
+
means = []
|
20
|
+
@arms.each do |arm|
|
21
|
+
ts["#{arm}:count"] = 0
|
22
|
+
ts["#{arm}:success"] = 0.0
|
23
|
+
end
|
24
|
+
ts.merge!(options)
|
25
|
+
|
26
|
+
@alpha, @beta = ts[:alpha].to_f, ts[:beta].to_f
|
27
|
+
means = []
|
28
|
+
|
29
|
+
@arms.each do |arm|
|
30
|
+
mean = self.class.beta_mean(
|
31
|
+
ts["#{arm}:success"],
|
32
|
+
ts["#{arm}:count"],
|
33
|
+
(ts["#{arm}:alpha"] || @alpha).to_f,
|
34
|
+
(ts["#{arm}:beta"] || @beta).to_f
|
35
|
+
)
|
36
|
+
means << [mean, arm]
|
37
|
+
end
|
38
|
+
|
39
|
+
@redis.multi do |r|
|
40
|
+
r.mapped_hmset @name, ts
|
41
|
+
r.sadd @arms_key, arms
|
42
|
+
r.zadd @means_key, means
|
43
|
+
end
|
44
|
+
self
|
45
|
+
end
|
46
|
+
|
47
|
+
def load!
|
48
|
+
@arms = Set.new @redis.smembers(@arms_key)
|
49
|
+
alpha, beta = @redis.hmget(@name, :alpha, :beta)
|
50
|
+
@alpha, @beta = alpha.to_f, beta.to_f
|
51
|
+
self
|
52
|
+
end
|
53
|
+
|
54
|
+
def delete!
|
55
|
+
@redis.multi do |r|
|
56
|
+
r.del @means_key
|
57
|
+
r.del @name
|
58
|
+
r.del @arms_key
|
59
|
+
end
|
60
|
+
end
|
61
|
+
|
62
|
+
def put(arm, options={})
|
63
|
+
ts = {"#{arm}:count" => 0, "#{arm}:success" => 0.0}
|
64
|
+
ts.merge! options
|
65
|
+
@arms << arm
|
66
|
+
@redis.multi do |r|
|
67
|
+
r.mapped_hmset @name, ts
|
68
|
+
r.sadd @arms_key, arm
|
69
|
+
end
|
70
|
+
update_mean arm
|
71
|
+
self
|
72
|
+
end
|
73
|
+
|
74
|
+
def remove(arm)
|
75
|
+
@arms.delete arm
|
76
|
+
@redis.multi do |r|
|
77
|
+
r.srem @arms_key, arm
|
78
|
+
r.hdel @name, ["#{arm}:success", "#{arm}:count", "#{arm}:alpha", "#{arm}:beta"]
|
79
|
+
r.zrem @means, arm
|
80
|
+
end
|
81
|
+
end
|
82
|
+
|
83
|
+
|
84
|
+
def disable(arm)
|
85
|
+
@redis.zrem @means, arm
|
86
|
+
end
|
87
|
+
|
88
|
+
def draw_multi(n)
|
89
|
+
drawn = []
|
90
|
+
n.times { drawn << draw }
|
91
|
+
drawn
|
92
|
+
end
|
93
|
+
|
94
|
+
def draw
|
95
|
+
max_arm = @redis.zrange(@means_key, -1, -1)[0]
|
96
|
+
@redis.hincrby @name, "#{max_arm}:count", 1
|
97
|
+
update_mean(max_arm)
|
98
|
+
max_arm
|
99
|
+
end
|
100
|
+
|
101
|
+
def update_success(arm, reward=1.0)
|
102
|
+
@redis.hincrbyfloat @name, "#{arm}:success", reward
|
103
|
+
update_mean(arm)
|
104
|
+
end
|
105
|
+
|
106
|
+
def update_mean(arm)
|
107
|
+
@redis.zadd @means_key, mean(arm), arm
|
108
|
+
end
|
109
|
+
alias enable update_mean
|
110
|
+
|
111
|
+
def mean(arm)
|
112
|
+
success, count, alpha, beta = @redis.hmget(@name, "#{arm}:success", "#{arm}:count", "#{arm}:alpha", "#{arm}:beta")
|
113
|
+
self.class.beta_mean success.to_f, count.to_f, (alpha || @alpha).to_f, (beta || @beta).to_f
|
114
|
+
end
|
115
|
+
|
116
|
+
def stats
|
117
|
+
arms, state, means = @redis.multi do |r|
|
118
|
+
r.smembers(@arms_key)
|
119
|
+
r.hgetall @name
|
120
|
+
r.zrange @means_key, 0, -1, :with_scores => true
|
121
|
+
end
|
122
|
+
|
123
|
+
{ :arms => arms, :state => state, :means => means }
|
124
|
+
end
|
125
|
+
end
|
126
|
+
end
|
@@ -0,0 +1,20 @@
|
|
1
|
+
# -*- encoding: utf-8 -*-
|
2
|
+
lib = File.expand_path('../lib', __FILE__)
|
3
|
+
$LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
|
4
|
+
require 'multi-armed-bandit/version'
|
5
|
+
|
6
|
+
Gem::Specification.new do |gem|
|
7
|
+
gem.name = "multi-armed-bandit"
|
8
|
+
gem.version = MultiArmedBandit::VERSION
|
9
|
+
gem.authors = ["David Dai"]
|
10
|
+
gem.email = ["ddai@scribd.com"]
|
11
|
+
gem.description = %q{A Redis backed multi-armed bandit library.}
|
12
|
+
gem.summary = %q{A Redis backed multi-armed bandit library.}
|
13
|
+
gem.homepage = "https://github.com/newtonapple/multi-armed-bandit"
|
14
|
+
|
15
|
+
gem.files = `git ls-files`.split($/)
|
16
|
+
gem.executables = gem.files.grep(%r{^bin/}).map{ |f| File.basename(f) }
|
17
|
+
gem.test_files = gem.files.grep(%r{^(test|spec|features)/})
|
18
|
+
gem.require_paths = ["lib"]
|
19
|
+
gem.add_dependency "redis", '~> 3.0.3'
|
20
|
+
end
|
metadata
ADDED
@@ -0,0 +1,66 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: multi-armed-bandit
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
version: 0.0.2
|
5
|
+
platform: ruby
|
6
|
+
authors:
|
7
|
+
- David Dai
|
8
|
+
autorequire:
|
9
|
+
bindir: bin
|
10
|
+
cert_chain: []
|
11
|
+
date: 2013-04-01 00:00:00.000000000 Z
|
12
|
+
dependencies:
|
13
|
+
- !ruby/object:Gem::Dependency
|
14
|
+
name: redis
|
15
|
+
requirement: !ruby/object:Gem::Requirement
|
16
|
+
requirements:
|
17
|
+
- - ~>
|
18
|
+
- !ruby/object:Gem::Version
|
19
|
+
version: 3.0.3
|
20
|
+
type: :runtime
|
21
|
+
prerelease: false
|
22
|
+
version_requirements: !ruby/object:Gem::Requirement
|
23
|
+
requirements:
|
24
|
+
- - ~>
|
25
|
+
- !ruby/object:Gem::Version
|
26
|
+
version: 3.0.3
|
27
|
+
description: A Redis backed multi-armed bandit library.
|
28
|
+
email:
|
29
|
+
- ddai@scribd.com
|
30
|
+
executables: []
|
31
|
+
extensions: []
|
32
|
+
extra_rdoc_files: []
|
33
|
+
files:
|
34
|
+
- .gitignore
|
35
|
+
- Gemfile
|
36
|
+
- LICENSE.txt
|
37
|
+
- README.md
|
38
|
+
- Rakefile
|
39
|
+
- lib/multi-armed-bandit.rb
|
40
|
+
- lib/multi-armed-bandit/thompson_sampling.rb
|
41
|
+
- lib/multi-armed-bandit/version.rb
|
42
|
+
- multi-armed-bandit.gemspec
|
43
|
+
homepage: https://github.com/newtonapple/multi-armed-bandit
|
44
|
+
licenses: []
|
45
|
+
metadata: {}
|
46
|
+
post_install_message:
|
47
|
+
rdoc_options: []
|
48
|
+
require_paths:
|
49
|
+
- lib
|
50
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
51
|
+
requirements:
|
52
|
+
- - '>='
|
53
|
+
- !ruby/object:Gem::Version
|
54
|
+
version: '0'
|
55
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
56
|
+
requirements:
|
57
|
+
- - '>='
|
58
|
+
- !ruby/object:Gem::Version
|
59
|
+
version: '0'
|
60
|
+
requirements: []
|
61
|
+
rubyforge_project:
|
62
|
+
rubygems_version: 2.0.0
|
63
|
+
signing_key:
|
64
|
+
specification_version: 4
|
65
|
+
summary: A Redis backed multi-armed bandit library.
|
66
|
+
test_files: []
|