Start : 28 Jun, 2018
- It is easy to learn algorithm while learning machine learning.
- I want to provide the easiest way to try making Machine Learning Algorithm.
- Need to develop the regression
- Making prototype of classifier is done.
- It deals all feature as categorical, not continuos.
- Entropy calcuation has contained. entropy of each feature is important to predicting.
- feature scaling (not need scaling when calculation entropy)
- calculate entropy (entropy is weight.)
- calculate feature weight (some feature is how many important than others)
- summary data point by level, stack to the recipe (the model)
- call predict function with test data
- get the suited point of every feature as test data's value from recipe (the making of Burger Matrix)
- product weight vector(in 3.) Burger Matrix(in 6.)
- find the best probable class.
See the practice file, know the using CheeseBurger Library simply.
/example/titanic-cheeseburger.py # the accuracy is 0.72952 (above 72%)
And compare with 'decision Tree' algorithm.
/example/titanic-decision-tree.py # the accuray is 0.77990 (above 77%)