Save models and RandomForestRegressor

[gcasadesus]

Description

1. Provides methods regarding the saving and loading of models using JSON and CBOR format.

   • save_model: Saves a serialized object to disk. This function uses the json and cbor2 utilities for serialization. Models, ds-arrays, and dictionaries of all kinds of objects can be saved using this function.
   • load_model: Uses a custom decoder to deserialize saved object files to memory.

2. Implemented RandomForestRegressor by modifying the already existing RandomForestClassifier. Mainly changed

   • the information of the leaf nodes from class frequency and mode to the mean value of the samples and
   • the splitting criterion from Gini Impurity (misclassification metric) to Mean Square Error (variance reduction metric).

3. Creation of a commons module that includes code used in multiple tasks to improve code maintainability. Following the already existing naming conventions, the rf submodule has been created with the classes regarding Random Forests and Decision Trees. The Random Forest Classifier and Regressor classes are exposed in the modules classification and regression, respectively, to allow backward compatibility.

Fixes #306

Type of change

• New algorithm or support class.
• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

To ensure correct functionality the saved and loaded models have been tested with the test cases already implemented in the library.

• I have added new test files.
• I have added new test cases.
• I have tested it manually in a local environment.
• I have tested it manually in a supercomputer.

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas.
• I have documented the public methods of any public class according to the guide styles.
• I have made corresponding changes to the documentation
• New and existing unit tests pass locally with my changes
• I have rebased my branch before trying to merge.

[codecov]

Codecov Report

Merging #349 (4a44cd3) into master (66e98b4) will increase coverage by 1.81%.
The diff coverage is 98.36%.

@@            Coverage Diff             @@
##           master     #349      +/-   ##
==========================================
+ Coverage   95.95%   97.77%   +1.81%     
==========================================
  Files          37       38       +1     
  Lines        3759     4001     +242     
==========================================
+ Hits         3607     3912     +305     
+ Misses        152       89      -63     

Impacted Files	Coverage Δ
dislib/classification/csvm/base.py	95.49% <ø> (ø)
dislib/commons/rf/forest.py	97.71% <97.71%> (ø)
dislib/commons/rf/decision_tree.py	99.01% <97.82%> (ø)
dislib/utils/saving.py	98.61% <98.61%> (ø)
dislib/commons/rf/data.py	98.76% <98.76%> (ø)
dislib/classification/__init__.py	100.00% <100.00%> (ø)
dislib/commons/rf/test_split.py	100.00% <100.00%> (ø)
dislib/regression/__init__.py	100.00% <100.00%> (ø)
dislib/utils/__init__.py	100.00% <100.00%> (ø)
... and 3 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 66e98b4...4a44cd3. Read the comment docs.