Notes

9.
from numpy import *
import operator
from os import listdir
import matplotlib
import matplotlib.pyplot as plt
import pandas as pd
from numpy.linalg import *

def kernel(point, xmat, k):
m, n = shape(xmat)
weights = mat(eye((m)))
for j in range(m):
diff = point - xmat[j]
weights[j, j] = exp(diff * diff.T / (-2.0 * k**2))
return weights

def localWeight(point, xmat, ymat, k):
wei = kernel(point, xmat, k)
W = (xmat.T * (wei * xmat)).I * (xmat.T * (wei * ymat.T))
return W

def localWeightRegression(xmat, ymat, k):
m, n = shape(xmat)
ypred = zeros(m)
for i in range(m):
ypred[i] = xmat[i] * localWeight(xmat[i], xmat, ymat, k)
return ypred

data = pd.read_csv('tips.csv')
bill = array(data.total_bill)
tip = array(data.tip)
mbill = mat(bill)
mtip = mat(tip)
m = shape(mbill)[1]
one = mat(ones(m))
X = hstack((one.T, mbill.T))

# set k here
ypred = localWeightRegression(X, mtip, 10)
SortIndex = X[:, 1].argsort(0)
xsort = X[SortIndex][:, 0]

fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
ax.scatter(bill, tip, color='green')
ax.plot(xsort[:, 1], ypred[SortIndex], color='red', linewidth=5)
plt.xlabel('Total bill')
plt.ylabel('Tip')
plt.show()

8.
from sklearn.datasets import load_iris
from sklearn.neighbors import KNeighborsClassifier
import numpy as np
from sklearn.model_selection import train_test_split

iris_dataset = load_iris()
print("n IRIS FEATURES TARGET NAMES: n ", iris_dataset.target_names)
for i in range(len(iris_dataset.target_names)):
print("n[{0}]:[{1}]".format(i, iris_dataset.target_names[i]))
print("n IRIS DATA :n", iris_dataset["data"])
X_train, X_test, y_train, y_test = train_test_split(iris_dataset["data"],
iris_dataset["target"], random_state=0)
print("n Target :n", iris_dataset["target"])
print("n X TRAIN n", X_train)
print("n X TEST n", X_test)
print("n Y TRAIN n", y_train)
print("n Y TEST n", y_test)

kn = KNeighborsClassifier(n_neighbors=1)
kn.fit(X_train, y_train)

for i in range(len(X_test)):
x = X_test[i]
x_new = np.array([x])
prediction = kn.predict(x_new)
print("n Actual : {0} {1}, Predicted :{2}{3}".format(
y_test[i], iris_dataset["target_names"][y_test[i]],
prediction, iris_dataset["target_names"][prediction]
))

print("n TEST SCORE[ACCURACY]: {:.2f}n".format(kn.score(X_test, y_test)))

7.
from copy import deepcopy
import numpy as np
import pandas as pd
from matplotlib import pyplot as plt
from sklearn.mixture import GaussianMixture
from sklearn.cluster import KMeans

data = pd.read_csv('ex.csv')
print("Input Data and Shape")
print(data.shape)
data.head()

print(data.head())
f1 = data['V1'].values
print("f1")
print(f1)
f2 = data['V2'].values
X = np.array(list(zip(f1, f2)))
print("x")
print(X)
print('Graph for the whole dataset')
plt.scatter(f1, f2, c='black', s=600)
plt.show()

kmeans = KMeans(2, random_state=0)
labels = kmeans.fit(X).predict(X)
print("labels")
print(labels)
centroids = kmeans.cluster_centers_
print("centroids")
print(centroids)
plt.scatter(X[:, 0], X[:, 1], c=labels, s=40)
print('Graph using Kmeans Algorithm')
plt.scatter(centroids[:, 0], centroids[:, 1], marker='*', s=200, c='#050505')
plt.show()

gmm = GaussianMixture(n_components=2).fit(X)
labels = gmm.predict(X)
print("lLABELS GMM")
print(labels)
probs = gmm.predict_proba(X)
size = 10 * probs.max(1) ** 3
print('Graph using EM Algorithm')
plt.scatter(X[:, 0], X[:, 1], c=labels, s=size, cmap='viridis')
plt.show()

6.
import numpy as np
import pandas as pd

mush = pd.read_csv("mushroom.csv")
mush.replace('?', np.nan, inplace=True)

print(len(mush.columns), "columns, after dropping NA,", len(mush.dropna(axis=1).columns))
mush.dropna(axis=1, inplace=True)

target = 'class'
features = mush.columns[mush.columns != target]
classes = mush[target].unique()
test = mush.sample(frac=0.3)
mush = mush.drop(test.index)

probs = {}
probcl = {}

for x in classes:
mushcl = mush[mush[target] == x][features]
clsp = {}
tot = len(mushcl)

for col in mushcl.columns:
colp = {}

for val, cnt in mushcl[col].value_counts().iteritems():
pr = cnt / tot
colp[val] = pr

clsp[col] = colp
probs[x] = clsp
probcl[x] = len(mushcl) / len(mush)


def probabs(x):
if not isinstance(x, pd.Series):
raise IOError("Arg must be of type Series")

probab = {}

for cl in classes:
pr = probcl[cl]

for col, val in x.iteritems():
try:
pr *= probs[cl][col][val]
except KeyError:
pr = 0

probab[cl] = pr

return probab


def classify(x):
probab = probabs(x)
mx = 0
mxcl = ''

for cl, pr in probab.items():
if pr > mx:
mx = pr
mxcl = cl

return mxcl


b = []

for i in mush.index:
b.append(classify(mush.loc[i, features]) == mush.loc[i, target])

print(sum(b), "correct of", len(mush))
print("Accuracy:", sum(b) / len(mush))

5.
import numpy as np

X = np.array(([2, 9], [1, 5], [3, 6]), dtype=float)
y = np.array(([92], [86], [89]), dtype=float)

X = X / np.amax(X, axis=0)
y = y / 100

def sigmoid(x):
return 1 / (1 + np.exp(-x))

def derivatives_sigmoid(x):
return x * (1 - x)

epoch = 7000
lr = 0.1
inputlayer_neurons = 2
hiddenlayer_neurons = 3
output_neurons = 1

wh = np.random.uniform(size=(inputlayer_neurons, hiddenlayer_neurons))
bh = np.random.uniform(size=(1, hiddenlayer_neurons))
wout = np.random.uniform(size=(hiddenlayer_neurons, output_neurons))
bout = np.random.uniform(size=(1, output_neurons))

for i in range(epoch):
# Forward Propagation
hinp1 = np.dot(X, wh)
hinp = hinp1 + bh
hlayer_act = sigmoid(hinp)

outinp1 = np.dot(hlayer_act, wout)
outinp = outinp1 + bout
output = sigmoid(outinp)

# Backpropagation
EO = y - output
outgrad = derivatives_sigmoid(output)
d_output = EO * outgrad

EH = d_output.dot(wout.T)
hiddengrad = derivatives_sigmoid(hlayer_act)
d_hiddenlayer = EH * hiddengrad

wout += hlayer_act.T.dot(d_output) * lr
wh += X.T.dot(d_hiddenlayer) * lr

print("Input: n" + str(X))
print("Actual Output: n" + str(y))
print("Predicted Output: n", output)

4.
import pandas as pd
import math
import numpy as np

data = pd.read_csv("tennis.csv")
features = [feat for feat in data]
features.remove("answer")


class Node:
def __init__(self):
self.children = []
self.value = ""
self.isLeaf = False
self.pred = ""


def entropy(examples):
pos = 0.0
neg = 0.0
for _, row in examples.iterrows():
if row["answer"] == "yes":
pos += 1
else:
neg += 1
if pos == 0.0 or neg == 0.0:
return 0.0
else:
p = pos / (pos + neg)
n = neg / (pos + neg)
return -(p * math.log(p, 2) + n * math.log(n, 2))


def info_gain(examples, attr):
uniq = np.unique(examples[attr])
gain = entropy(examples)
for u in uniq:
subdata = examples[examples[attr] == u]
sub_e = entropy(subdata)
gain -= (float(len(subdata)) / float(len(examples))) * sub_e
return gain


def ID3(examples, attrs):
root = Node()
max_gain = 0
max_feat = ""
for feature in attrs:
gain = info_gain(examples, feature)
if gain > max_gain:
max_gain = gain
max_feat = feature
root.value = max_feat
uniq = np.unique(examples[max_feat])
for u in uniq:
subdata = examples[examples[max_feat] == u]
if entropy(subdata) == 0.0:
newNode = Node()
newNode.isLeaf = True
newNode.value = u
newNode.pred = np.unique(subdata["answer"])
root.children.append(newNode)
else:
dummyNode = Node()
dummyNode.value = u
new_attrs = attrs.copy()
new_attrs.remove(max_feat)
child = ID3(subdata, new_attrs)
dummyNode.children.append(child)
root.children.append(dummyNode)
return root


def printTree(root: Node, depth=0):
for i in range(depth):
print("t", end="")
print(root.value, end="")
if root.isLeaf:
print(" -> ", root.pred)
print()
for child in root.children:
printTree(child, depth + 1)


def classify(root: Node, new):
for child in root.children:
if child.value == new[root.value]:
if child.isLeaf:
print("Predicted Label for new example", new, " is:", child.pred)
exit
else:
classify(child.children[0], new)


root = ID3(data, features)
print("Decision Tree is:")
printTree(root)
print("------------------")
new = {"outlook": "sunny", "temperature": "hot", "humidity": "normal", "wind":
"strong"}
classify(root, new)

3.
import numpy as np
import pandas as pd

data = pd.DataFrame(data=pd.read_csv('finds.csv'))
concepts = np.array(data.iloc[:, 0:-1])
target = np.array(data.iloc[:, -1])


def learn(concepts, target):
specific_h = concepts[0].copy()
general_h = [["?" for i in range(len(specific_h))] for i in range(len(specific_h))]

for i, h in enumerate(concepts):
if target[i] == "Yes":
for x in range(len(specific_h)):
if h[x] != specific_h[x]:
specific_h[x] = '?'
general_h[x][x] = '?'
if target[i] == "No":
for x in range(len(specific_h)):
if h[x] != specific_h[x]:
general_h[x][x] = specific_h[x]
else:
general_h[x][x] = '?'

indices = [i for i, val in enumerate(general_h) if val == ['?', '?', '?', '?', '?', '?']]
for i in indices:
general_h.remove(['?', '?', '?', '?', '?', '?'])

return specific_h, general_h


s_final, g_final = learn(concepts, target)
print("Final S:", s_final, sep="n")
print("Final G:", g_final, sep="n")

2.
class Graph:
def __init__(self, adjac_lis):
self.adjac_lis = adjac_lis
self.H1 = {
'A': 1,
'B': 6,
'C': 2,
'D': 2,
'E': 2,
'F': 1,
'G': 5,
'H': 7,
'I': 7,
'J': 1,
'T': 3
}
self.H = {
'A': 1,
'B': 6,
'C': 12,
'D': 10,
'E': 4,
'F': 4,
'G': 5,
'H': 7,
}
self.parent = {}
self.openList = set()
self.hasRevised = []
self.solutionGraph = {}
self.solvedNodeList = {}

def get_neighbors(self, v):
return self.adjac_lis.get(v, '')

def updateNode(self, v):
if v in self.solvedNodeList:
return

feasibleChildNodeList = []
minimumCost = None
minimumCostFeasibleChildNodesDict = {}

print("CURRENT PROCESSING NODE:", v)
print("___________________________")

for (c, weight) in self.get_neighbors(v):
feasibleChildNodeList = []
cost = self.getHeuristicNodeValue(c) + 1
feasibleChildNodeList.append(c)

andNodesList = self.getAndNodes(v)
for nodeTuple in andNodesList:
if c in nodeTuple:
for andNode in nodeTuple:
if andNode != c:
feasibleChildNodeList.append(andNode)
cost = cost + self.getHeuristicNodeValue(andNode) + 1

if minimumCost == None:
minimumCost = cost
for child in feasibleChildNodeList:
self.parent[child] = v
minimumCostFeasibleChildNodesDict[minimumCost] = feasibleChildNodeList
else:
if minimumCost > cost:
minimumCost = cost
for child in feasibleChildNodeList:
self.parent[child] = v
minimumCostFeasibleChildNodesDict[minimumCost] = feasibleChildNodeList

if minimumCost == None:
minimumCost = self.getHeuristicNodeValue(v)
self.solvedNodeList.add(v)
else:
self.setHeuristicNodeValue(v, minimumCost)

for child in minimumCostFeasibleChildNodesDict[minimumCost]:
if child not in self.solvedNodeList:
self.openList.add(child)
self.solutionGraph[v] = minimumCostFeasibleChildNodesDict[minimumCost]

solved = True
for c in self.solutionGraph.get(v, ''):
if c not in self.solvedNodeList:
solved = solved & False

if solved == True:
self.solvedNodeList.add(v)

print("HEURISTIC VALUES :", self.H)
print("OPEN LIST :", list(self.openList))
print("MINIMUM COST NODES:", minimumCostFeasibleChildNodesDict.get(minimumCost, "[ ]"))
print("SOLVED NODE LIST :", list(self.solvedNodeList))
print("-----------------------------------------------------------------------------------------")

def getAndNodes(self, v):
andNodes = {
'A': [('B', 'C')],
'D': [('E', 'F')]
}
return andNodes.get(v, '')

def getHeuristicNodeValue(self, n):
return self.H.get(n, 0)

def setHeuristicNodeValue(self, n, value):
self.H[n] = value

def ao_star_algorithm(self, start):
self.openList = set([start])

while len(self.openList) > 0:
v = self.openList.pop()
self.updateNode(v)

while v != start and self.parent[v] not in self.solvedNodeList:
parent = self.parent[v]
self.updateNode(parent)
v = parent

print("TRAVERSE SOLUTION FROM ROOT TO COMPUTE THE FINAL SOLUTION GRAPH")
print("---------------------------------------------------------------")
print("SOLUTION GRAPH:", self.solutionGraph)
print("n")


nodeList1 = {
'A': [('B', 1), ('C', 1), ('D', 1)],
'B': [('G', 1), ('H', 1)],
'C': [('J', 1)],
'D': [('E', 1), ('F', 1)],
'G': [('I', 1)]
}

nodeList = {
'A': [('B', 1), ('C', 1), ('D', 1)],
'B': [('G', 1), ('H', 1)],
'D': [('E', 1), ('F', 1)]
}

graph = Graph(nodeList)
graph.ao_star_algorithm('A')

1.
from collections import deque

class Graph:
def __init__(self, adjac_lis):
self.adjac_lis = adjac_lis

def get_neighbors(self, v):
return self.adjac_lis[v]

def h(self, n):
H = {
'A': 1,
'B': 1,
'C': 1,
'D': 1
}
return H[n]

def a_star_algorithm(self, start, stop):
open_lst = set([start])
closed_lst = set([])
poo = {}
poo[start] = 0
par = {}
par[start] = start

while len(open_lst) > 0:
n = None
for v in open_lst:
if n is None or poo[v] + self.h(v) < poo[n] + self.h(n):
n = v

if n is None:
print('Path does not exist!')
return None

if n == stop:
reconst_path = []
while par[n] != n:
reconst_path.append(n)
n = par[n]
reconst_path.append(start)
reconst_path.reverse()
print('Path found: {}'.format(reconst_path))
return reconst_path

for (m, weight) in self.get_neighbors(n):
if m not in open_lst and m not in closed_lst:
open_lst.add(m)
par[m] = n
poo[m] = poo[n] + weight
else:
if poo[m] > poo[n] + weight:
poo[m] = poo[n] + weight
par[m] = n

if m in closed_lst:
closed_lst.remove(m)
open_lst.add(m)

open_lst.remove(n)
closed_lst.add(n)

print('Path does not exist!')
return None