Implement functionality for basic decision tree (classification).

[gpestre]

Absorbs most of the functions from helpers.cpp:

• [x] Representing data:

datasets.cpp : DataFrame https://github.com/johannes-kk/cs205_final_project/blob/718cc1acc9c908f888ccbdd9acea6afeab500411/src/helpers.cpp#L13

datasets.cpp : DataVector https://github.com/johannes-kk/cs205_final_project/blob/718cc1acc9c908f888ccbdd9acea6afeab500411/src/helpers.cpp#L14

• [x] Calculating losses:

losses.cpp : LossFunction("gini_impurity") : https://github.com/johannes-kk/cs205_final_project/blob/718cc1acc9c908f888ccbdd9acea6afeab500411/src/helpers.cpp#L80

losses.cpp : LossFunction("misclassification_rate") : https://github.com/johannes-kk/cs205_final_project/blob/718cc1acc9c908f888ccbdd9acea6afeab500411/src/helpers.cpp#L20

• [x] Splitting datasets:

datasets.cpp :

// Returns a pair of frames (value above and below threshold in specified column).
std::vector<DataFrame*> split(int split_column, double split_threshold, bool equal_goes_left) const; 

https://github.com/johannes-kk/cs205_final_project/blob/718cc1acc9c908f888ccbdd9acea6afeab500411/src/helpers.cpp#L37

• [x] Finding best split:

losses.cpp ? : https://github.com/johannes-kk/cs205_final_project/blob/718cc1acc9c908f888ccbdd9acea6afeab500411/src/helpers.cpp#L113

[johannes-kk]

Amazing! 🚀

[gpestre]

Doesn't implement prediction yet, but has a fit method that seems to run without error. Definitely needs some stress-testing though.