added some comments
This commit is contained in:
parent
7e92796a19
commit
26a2824818
86
main.py
86
main.py
@ -56,45 +56,45 @@ tree_format = ["dish", "temperature", "label"]
|
|||||||
|
|
||||||
#print(func_output)
|
#print(func_output)
|
||||||
|
|
||||||
def uniq_val_from_data(rows, col):
|
def uniq_count(rows):
|
||||||
return set([row[col] for row in rows])
|
#count uniq labels(names)
|
||||||
|
count = {}
|
||||||
|
|
||||||
def class_counts(rows):
|
|
||||||
counts = {}
|
|
||||||
for row in rows:
|
for row in rows:
|
||||||
label = row[-1]
|
lbl = row[-1]
|
||||||
if label not in counts:
|
if lbl not in count:
|
||||||
counts[label] = 0
|
count[lbl] = 0
|
||||||
counts[label] += 1
|
count[lbl] += 1
|
||||||
return counts
|
return count
|
||||||
|
|
||||||
|
#didn't used
|
||||||
def isnumer(value):
|
def isnumer(val):
|
||||||
return isinstance(value, int) or isinstance(value, float)
|
return isinstance(val, int) or isinstance(val, float)
|
||||||
|
|
||||||
|
|
||||||
class Question():
|
class Question():
|
||||||
|
|
||||||
def __init__(self, col, value):
|
def __init__(self, col, value):
|
||||||
self.col = col
|
self.col = col #column
|
||||||
self.value = value
|
self.value = value #value of column
|
||||||
|
|
||||||
def compare(self, example):
|
def compare(self, example):
|
||||||
|
#compare val in example with val in the question
|
||||||
val = example[self.col]
|
val = example[self.col]
|
||||||
if isnumer(val):
|
if isnumer(val): #in case menu have prices
|
||||||
return val >= self.value
|
return val >= self.value
|
||||||
else:
|
else:
|
||||||
return val == self.value
|
return val == self.value
|
||||||
|
|
||||||
def __repr__(self):
|
def __repr__(self):
|
||||||
|
#just to print
|
||||||
condition = "=="
|
condition = "=="
|
||||||
if isnumer(self.value):
|
if isnumer(self.value):
|
||||||
condition = ">="
|
condition = ">="
|
||||||
return "Is %s %s %s?" % (tree_format[self.col], condition, str(self.value))
|
return "Is %s %s %s?" % (tree_format[self.col], condition, str(self.value))
|
||||||
|
|
||||||
|
|
||||||
def partition(rows, quest):
|
def split(rows, quest):
|
||||||
|
#split data into True and False
|
||||||
t_rows, f_rows = [], []
|
t_rows, f_rows = [], []
|
||||||
for row in rows:
|
for row in rows:
|
||||||
if quest.compare(row):
|
if quest.compare(row):
|
||||||
@ -105,7 +105,7 @@ def partition(rows, quest):
|
|||||||
|
|
||||||
|
|
||||||
def gini(rows):
|
def gini(rows):
|
||||||
counts = class_counts(rows)
|
counts = uniq_count(rows)
|
||||||
impurity = 1
|
impurity = 1
|
||||||
for lbl in counts:
|
for lbl in counts:
|
||||||
prob_of_lbl = counts[lbl] / float(len(rows))
|
prob_of_lbl = counts[lbl] / float(len(rows))
|
||||||
@ -114,11 +114,12 @@ def gini(rows):
|
|||||||
|
|
||||||
|
|
||||||
def info_gain(l, r, current_uncertainty):
|
def info_gain(l, r, current_uncertainty):
|
||||||
p = float(len(l)) / (len(l) + len(r))
|
p = float(len(l)) / (len(l) + len(r)) #something like an enthropy?
|
||||||
return current_uncertainty - p*gini(l) - (1-p)*gini(r)
|
return current_uncertainty - p*gini(l) - (1-p)*gini(r)
|
||||||
|
|
||||||
|
|
||||||
def find_best_q(rows):
|
def find_best_q(rows):
|
||||||
|
#best question to split the data
|
||||||
best_gain = 0
|
best_gain = 0
|
||||||
best_quest = None
|
best_quest = None
|
||||||
current_uncertainty = gini(rows)
|
current_uncertainty = gini(rows)
|
||||||
@ -130,7 +131,7 @@ def find_best_q(rows):
|
|||||||
for val in vals:
|
for val in vals:
|
||||||
quest = Question(col, val)
|
quest = Question(col, val)
|
||||||
|
|
||||||
t_rows, f_rows = partition(rows, quest)
|
t_rows, f_rows = split(rows, quest)
|
||||||
|
|
||||||
if len(t_rows) == 0 or len(f_rows) == 0:
|
if len(t_rows) == 0 or len(f_rows) == 0:
|
||||||
continue
|
continue
|
||||||
@ -144,11 +145,13 @@ def find_best_q(rows):
|
|||||||
|
|
||||||
|
|
||||||
class Leaf:
|
class Leaf:
|
||||||
|
#contain a number of how many times the label has appeared in dataset
|
||||||
def __init__(self, rows):
|
def __init__(self, rows):
|
||||||
self.predicts = class_counts(rows)
|
self.predicts = uniq_count(rows)
|
||||||
|
|
||||||
|
|
||||||
class Decision_Node():
|
class Decision_Node():
|
||||||
|
#contain the question and child nodes
|
||||||
def __init__(self, quest, t_branch, f_branch):
|
def __init__(self, quest, t_branch, f_branch):
|
||||||
self.quest = quest
|
self.quest = quest
|
||||||
self.t_branch = t_branch
|
self.t_branch = t_branch
|
||||||
@ -156,37 +159,45 @@ class Decision_Node():
|
|||||||
|
|
||||||
|
|
||||||
def build_tree(rows):
|
def build_tree(rows):
|
||||||
|
#use info gain and question
|
||||||
gain, quest = find_best_q(rows)
|
gain, quest = find_best_q(rows)
|
||||||
|
|
||||||
|
#no gain = no more question, so return a Leaf
|
||||||
if gain == 0:
|
if gain == 0:
|
||||||
return Leaf(rows)
|
return Leaf(rows)
|
||||||
|
|
||||||
t_rows, f_rows = partition(rows, quest)
|
#split into true and false branch
|
||||||
|
t_rows, f_rows = split(rows, quest)
|
||||||
|
|
||||||
|
#print out branches
|
||||||
t_branch = build_tree(t_rows)
|
t_branch = build_tree(t_rows)
|
||||||
f_branch = build_tree(f_rows)
|
f_branch = build_tree(f_rows)
|
||||||
|
|
||||||
|
#return the child/leaf
|
||||||
return Decision_Node(quest, t_branch, f_branch)
|
return Decision_Node(quest, t_branch, f_branch)
|
||||||
|
|
||||||
|
|
||||||
def print_tree(node, spc=""):
|
def print_tree(node, spc=""):
|
||||||
|
|
||||||
|
#if node is a leaf
|
||||||
if isinstance(node, Leaf):
|
if isinstance(node, Leaf):
|
||||||
print(" " + "Predict", node.predicts)
|
print(" " + "Predict", node.predicts)
|
||||||
return
|
return #end of function
|
||||||
|
|
||||||
|
#Or question
|
||||||
print("" + str(node.quest))
|
print("" + str(node.quest))
|
||||||
|
#True branch
|
||||||
print("" + '--> True:')
|
print("" + '--> True:')
|
||||||
print_tree(node.t_branch, spc + " ")
|
print_tree(node.t_branch, spc + " ")
|
||||||
|
#False branch
|
||||||
print("" + '--> False:')
|
print("" + '--> False:')
|
||||||
print_tree(node.f_branch, spc + " ")
|
print_tree(node.f_branch, spc + " ")
|
||||||
|
|
||||||
def classify(row, node):
|
def classify(row, node):
|
||||||
|
#return our prediction in case the node is a leaf
|
||||||
if isinstance(node, Leaf):
|
if isinstance(node, Leaf):
|
||||||
return node.predictions
|
return node.predicts
|
||||||
|
#otherwise go to the child
|
||||||
if node.quest.compare(row):
|
if node.quest.compare(row):
|
||||||
return classify(row, node.t_branch)
|
return classify(row, node.t_branch)
|
||||||
else:
|
else:
|
||||||
@ -194,8 +205,9 @@ def classify(row, node):
|
|||||||
|
|
||||||
|
|
||||||
def print_leaf(counts):
|
def print_leaf(counts):
|
||||||
|
#count prediction
|
||||||
total = sum(counts.values())*1.0
|
total = sum(counts.values())*1.0
|
||||||
probs = {}
|
probs = {} #probability
|
||||||
for lbl in counts.keys():
|
for lbl in counts.keys():
|
||||||
probs[lbl] = str(int(counts[lbl] / total*100)) + "%"
|
probs[lbl] = str(int(counts[lbl] / total*100)) + "%"
|
||||||
return probs
|
return probs
|
||||||
|
Loading…
Reference in New Issue
Block a user