vamc-stash

import numpy as np 
import pandas as pd 
import matplotlib.pyplot as plt 
import math 
import operator


eps = np.finfo(float).eps

def rmse_score(y_true, y_pred):

vamc-stash

def _predict_target(self, feature_lookup, x, tree):

	for node in tree.keys():
		val = x[node]
		if type(val) == str:
			tree = tree[node][val]
		else:
			cutoff = str(list(tree[node].keys())[0]).split('<=')[1]

			if(val <= float(cutoff)):	#Left Child

vamc-stash

def _build_tree(self, df, tree = None):

  """
    Args:
      df: current number of rows available for splitting(decision making)

  """

  #Get feature with minimum score
  feature, cutoff = self._find_best_split(df)

vamc-stash

def _find_feature_split(self, feature, df):

  best_score = float('inf')
  cutoff = float

  for val in df[feature]:
    left_child = df[feature][df[feature] <= val] 
    right_child = df[feature][df[feature] > val]

    if(len(left_child) > 0 and len(right_child) > 0):

vamc-stash

def _find_best_split(self, df):

  """
    Finds the column to split on first.

  """

  best_feature = str
  cutoff = None
  best_score = float('inf')

vamc-stash

import numpy as np 
import pandas as pd 
import matplotlib.pyplot as plt 
import math 
import operator


eps = np.finfo(float).eps

def train_test_split(x, y, test_size = 0.25, random_state = None):

vamc-stash

import numpy as np 
import pandas as pd 
import matplotlib.pyplot as plt 
import math 
import operator


eps = np.finfo(float).eps

def accuracy_score(y_true, y_pred):

vamc-stash

def _predict_target(self, feature_lookup, x, tree):

	for node in tree.keys():
		val = x[node]
		if type(val) == str:
			tree = tree[node][val]
		else:
			cutoff = str(list(tree[node].keys())[0]).split('<=')[1]

			if(val <= float(cutoff)):	#Left Child

vamc-stash

def _build_tree(self, df, tree = None):

  """
    Args:
      df: current number of rows available for splitting(decision making)

  """

  #Get feature with maximum information gain
  feature, cutoff = self._find_best_split(df)

vamc-stash

def _find_best_split(self, df):

	"""
		Finds the column to split on first using 'Information Gain' Metric.

		Information Gain(IG) = Entropy(parent) - Sum of Entropy(Children)
					IG(T, a) = H(T) - H(T|a)

		Entropy(parent) H(T) = (Sum[i=1 to J](- Pi * log(Pi)))
		Sum of Entropy(children) H(T|a) = Sum(P(a) * Sum[i=1 to J](- P(i|a) * log(P(i|a)))