libmf 0.1.2 → 0.1.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 907cba459d6d4af0607aa2662b0f32b5afc30a1940cbf6c4e7712553cb95709c
-  data.tar.gz: 16e84358005ebb5432a2765c00dd79fe26f9822cbdfcb0570d829099c11e5470
+  metadata.gz: 316a859127d3ee4a6b2af41599daf5d14e1f436479dfd8d1c8ebd739b8141367
+  data.tar.gz: 59b64d67f90955b81630873bc2b776cbf5a463174502ccd3f9e92eb29810eac9
 SHA512:
-  metadata.gz: 56e8dd57eb146cf81ffd11c193565e632f623d0ade5c2650344e19951502bd492becf6268066da6c630936e5bb08e3ee068d9ecb7f942457955c88b1faf9e5eb
-  data.tar.gz: 2764dab3e8d5d3044bf78482988addeb40c026b592cb57ccec3984f70244eeaf77a0a8a4e597a8607d9cf41c74e69ab47391f1003ba163a45b84b010821a76c5
+  metadata.gz: 8b2d80a014a92dd78533e31476909aac348906ab0caba74d30826e9f7057bb5f9dd649f10c49d2d436325dbdedc8b590ab031c65197cc4c1b8be6290f525ace9
+  data.tar.gz: 9af9ef4372b7ed124bc3cbdc823e3dfbfc1e10303c914895fbaa608f5959a6de560349148cc6a83e35ef002409042705af198d5f533fbc0e3a87c6fbce0a438b

data/CHANGELOG.md

@@ -1,3 +1,8 @@
+## 0.1.3
+
+- Made parameter names more Ruby-like
+- No need to set `do_nmf` with generalized KL-divergence
+
 ## 0.1.2
 
 - Fixed bug in `p_factors` and `q_factors` methods

data/README.md

@@ -2,8 +2,6 @@
 
 [LIBMF](https://github.com/cjlin1/libmf) - large-scale sparse matrix factorization - for Ruby
 
-:fire: Uses the C API for blazing performance
-
 [![Build Status](https://travis-ci.org/ankane/libmf.svg?branch=master)](https://travis-ci.org/ankane/libmf)
 
 ## Installation
@@ -65,36 +63,6 @@ Pass a validation set
 model.fit(data, eval_set: eval_set)
 ```
 
-## Parameters
-
-Pass parameters
-
-```ruby
-model = Libmf::Model.new(k: 20, nr_iters: 50)
-```
-
-Supports the same parameters as LIBMF
-
-```text
-variable      meaning                                    default
-================================================================
-fun           loss function                                    0
-k             number of latent factors                         8
-nr_threads    number of threads used                          12
-nr_bins       number of bins                                  25
-nr_iters      number of iterations                            20
-lambda_p1     coefficient of L1-norm regularization on P       0
-lambda_p2     coefficient of L2-norm regularization on P     0.1
-lambda_q1     coefficient of L1-norm regularization on Q       0
-lambda_q2     coefficient of L2-norm regularization on Q     0.1
-eta           learning rate                                  0.1
-alpha         importance of negative entries                 0.1
-c             desired value of negative entries           0.0001
-do_nmf        perform non-negative MF (NMF)                false
-quiet         no outputs to stdout                         false
-copy_data     copy data in training procedure               true
-```
-
 ## Cross-Validation
 
 Perform cross-validation
@@ -109,6 +77,50 @@ Specify the number of folds
 model.cv(data, folds: 5)
 ```
 
+## Parameters
+
+Pass parameters - default values below
+
+```ruby
+Libmf::Model.new(
+  loss: 0,                # loss function
+  factors: 8,             # number of latent factors
+  threads: 12,            # number of threads used
+  bins: 25,               # number of bins
+  iterations: 20,         # number of iterations
+  lambda_p1: 0,           # coefficient of L1-norm regularization on P
+  lambda_p2: 0.1,         # coefficient of L2-norm regularization on P
+  lambda_q1: 0,           # coefficient of L1-norm regularization on Q
+  lambda_q2: 0.1,         # coefficient of L2-norm regularization on Q
+  learning_rate: 0.1,     # learning rate
+  alpha: 0.1,             # importance of negative entries
+  c: 0.0001,              # desired value of negative entries
+  nmf: false,             # perform non-negative MF (NMF)
+  quiet: false,           # no outputs to stdout
+  copy_data: true         # copy data in training procedure
+)
+```
+
+### Loss Functions
+
+For real-valued matrix factorization
+
+- 0 - squared error (L2-norm)
+- 1 - absolute error (L1-norm)
+- 2 - generalized KL-divergence
+
+For binary matrix factorization
+
+- 5 - logarithmic error
+- 6 - squared hinge loss
+- 7 - hinge loss
+
+For one-class matrix factorization
+
+- 10 - row-oriented pair-wise logarithmic loss
+- 11 - column-oriented pair-wise logarithmic loss
+- 12 - squared error (L2-norm)
+
 ## Resources
 
 - [LIBMF: A Library for Parallel Matrix Factorization in Shared-memory Systems](https://www.csie.ntu.edu.tw/~cjlin/papers/libmf/libmf_open_source.pdf)

data/lib/libmf/model.rb

@@ -68,11 +68,24 @@ module Libmf
 
     def param
       param = FFI.mf_get_default_param
+      options = @options.dup
       # silence insufficient blocks warning with default params
-      options = {nr_bins: 25}.merge(@options)
+      options[:bins] ||= 25 unless options[:nr_bins]
+      options_map = {
+        :loss => :fun,
+        :factors => :k,
+        :threads => :nr_threads,
+        :bins => :nr_bins,
+        :iterations => :nr_iters,
+        :learning_rate => :eta,
+        :nmf => :do_nmf
+      }
       options.each do |k, v|
+        k = options_map[k] if options_map[k]
         param[k] = v
       end
+      # do_nmf must be true for generalized KL-divergence
+      param[:do_nmf] = true if param[:fun] == 2
       param
     end
 

data/lib/libmf/version.rb

@@ -1,3 +1,3 @@
 module Libmf
-  VERSION = "0.1.2"
+  VERSION = "0.1.3"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: libmf
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.1.3
 platform: ruby
 authors:
 - Andrew Kane
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-11-07 00:00:00.000000000 Z
+date: 2019-11-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi