The document discusses a machine learning algorithm to perform classification on a diabetes dataset. It loads and splits the dataset into training and test sets. It then calculates the mean and standard deviation for each attribute within each class in the training set. These statistics are used to predict the class of each test instance and calculate the overall accuracy.

1. import csv, random, math # import csv,random,math module
import statistics as st # import statistics module
from statistics import stdev # from statistics library import
stdev method
1

2. # Loading the Data from csv file of
pima-indians-diabetes dataset.
def loadCsv(filename):
lines = csv.reader(open(filename, "r"));
#open csv file in read mode and read file
dataset = list(lines)
for i in range(len(dataset)):
# converting the attributes from string to floating point numbers
dataset[i] = [float(x) for x in dataset[i]]
return dataset
2

3. Note:
• dataset contains 768 observations of medical details for
Pima Indian patients.
• The attributes are :
1. Pregnancies
2. Glucose
3. Blood pressure
4. Skin thickness
5. Insulin
6. BMI
7. Diabetic pedigree function
8. Age
Outcome(diabetes) 1 or 0
3

4. def splitDataset(dataset, splitRatio):
testSize = int(len(dataset) * splitRatio); (768*0.2)=153
trainSet = list(dataset); // All 768 examples
testSet = [] // testSet is initially empty
while len(testSet) < testSize: #untill len(testSet) becomes 153
#randomly pick an instance from training data
index = random.randrange(len(trainSet));
testSet.append(trainSet.pop(index))
return [trainSet, testSet]
#Function will split dataset into trainingSet and testSet
#trainSet contains initially 768 examples
#randomly pick 153 instances from trainSet, remove it and
place it in testSet
i.e trainSet+testSet=768
4

5. def separateByClass(dataset):
separated = {}
for i in range(len(dataset)):
x = dataset[i] # x contains row of dataset
if (x[-1] not in separated):
separated[x[-1]] = []
separated[x[-1]].append(x)
return separated
#Function separateByClass groups training set into 2 classes.
1 group contains all positive instances.
1 group contains all negative instances.
#separated= dictionary, key 1 value= all positive classification
instances, key=0, value= all negative classification instances
• #Function to categorize the dataset in terms of classes (1 and 0)
• #The function assumes that the last attribute (-1) is the class value.
• x[-1]=1 or 0
5

6. 6
separated is a dictionary with key = 1 and 0
When key=1, values are all positive instances
When key=0, values are all negative instances

7. 7
separated is a dictionary with key values 1 and 0
When key=1, values are all positive instances
When key=0, values are all negative instances

8. def mean(numbers):
return sum(numbers)/float(len(numbers))
# Function to calculate mean
8

9. def stdev(numbers):
avg=mean(numbers)
variance=sum([pow(x-avg,2) for x in
numbers])/float(len(numbers)-1)
return math.sqrt(variance)
# Function to calculate standard deviation
9
# square root of variance is standard deviation

10. def compute_mean_std(dataset):
mean_std = [(st.mean(attribute), st.stdev(attribute))for attribute in
zip(*dataset)];
del mean_std[-1] # Exclude mean and standard deviation of last
column of dataset (i.e label 1 and 0)
return mean_std
# All class 1 data are zipped. For all 8 attributes calculate mean and
standard deviation.
# All class 0 data are zipped. For all 8 attributes calculate mean and
standard deviation.
# the function compute_mean computes the mean and standard
deviation of 8 attributes for both positive and negative instances
10

11. Mean, standard deviation of 8 attributes for class 1 and 0
class 1 data mean and standard deviation
class 0 data mean and standard deviation
11
1. Pregnancies
2. Glucose
3. Blood pressure
4. Skin thickness
5. Insulin
6. BMI
7. Diabetic pedigree function
8. Age

12. def summarizeByClass(dataset): dataset=TrainingSet
separated = separateByClass(dataset)
#The function separateByClass(dataset) will divide training set into 2
classes 1).All instances with class 1 (Having diabetes) 2) All
instances with class 0 (Not having diabetes)
12

13. #separated= dictionary, When key=0 value= all
positive classification instances, When key=1,
value= all negative classification instances
summary = {} # to store mean and standard
deviation of +ve and -ve instances
for classValue, instances in separated.items():
summary[classValue] = compute_mean_std(instances)
return summary
#summary is a dictionary of tuples(mean,std) for each
class value 13

14. for classValue, instances in separated.items():
#summary is a dictionary of tuples(mean,std) for each class value
summary[classValue] = compute_mean_std(instances)
return summary
classValue =1, instances= all positive class instances
classValue=0, instances=all negative class instances
14
separated is a dictionary with key values 1 and 0
When key=1, values are all positive instances
When key=0, values are all negative instances

15. summary[classValue] = compute_mean_std(instances)
summary[1]=mean and standard deviation of all positive instances
summary[0]=mean and standard deviation of all negative instances
15

16. def estimateProbability(x, mean, stdev):
exponent = math.exp(-(math.pow(x-mean,2)/(2*math.pow(stdev,2))))
return (1 / (math.sqrt(2*math.pi) * stdev)) * exponent
# Calculate probability density function
16

17. def calculateClassProbabilities(summaries, testVector):
p = {} testinstance
#class and attribute information as mean and standard deviation
for classValue, classSummaries in summaries.items():
p[classValue] = 1 #Initially p[1]=1 and p[0]=0
for i in range(len(classSummaries)):
mean, stdev = classSummaries[i]
x = testVector[i] #testvector's first attribute
p[classValue] *= estimateProbability(x, mean, stdev);
return p
#classValue will be 1 and 0
#classSummaries contains mean and standard deviation of each
class
#summaries contains mean and standard deviation of 8 attributes of
2 classes. 17

18. for i in range(len(classSummaries)):
mean, stdev = classSummaries[i]
18

19. x = testVector[i] #testSet’s first attribute
p[classValue] *= estimateProbability(x, mean, stdev);
19

20. Function look for the largest probability and return the associated class
def predict(summaries, testVector):
testVector= testSet instance i.e each row of testSet
all_p = calculateClassProbabilities(summaries, testVector)
#return probabilities of each class (1 and 0)
20

21. bestLabel, bestProb = None, -1
#initially bestLabel=None, bestProb=-1
for lbl, p in all_p.items():
# lbl will contain 1 and 0
#p contains probability value
if bestLabel is None or p > bestProb:
bestProb = p
bestLabel = lbl
return bestLabel
# compare probability values and return class label
with highest probability
21

22. if bestLabel is None or p > bestProb:
bestProb = p
bestLabel = lbl
return bestLabel
# Compare 2 probability values and return label
associated with highest probability
22

23. def perform_classification(summaries, testSet):
predictions = []
for i in range(len(testSet)):
result = predict(summaries, testSet[i])
predictions.append(result)
return predictions
// Function perform_classification return predictions
of all testSets (153)
23

24. def perform_classification(summaries, testSet):
predictions = [] # predictions list will store predicted
values of 153 testsets
# initially predictions list is empty
summaries contains mean and standard deviation of
8 attributes of 2 classes.
Class 1
Class 0
24

25. for i in range(len(testSet)): # for each testSet instance
result = predict(summaries, testSet[i])
#calculate prediction (1 or 0) and store in result
predictions.append(result)
#append each prediction result to list predictions
return predictions
25

26. # Function which return predictions for list of predictions #
For each instance
def getAccuracy(testSet, predictions):
correct = 0
for i in range(len(testSet)):
if testSet[i][-1] == predictions[i]:
correct += 1
return(correct/float(len(testSet))) * 100.0
Function takes 2 arguments testSet and predictions
• If testSet classification value matches with predictions
value of that testSet, then prediction is correct. Then
increment correct by 1.
• Accuracy=how many testSets are correctly predicted by
model.
• Accuracy=(no of correct predictions/153)*100 26

27. dataset = loadCsv('pima-indians-diabetes.csv');
#loadCsv function is called and entire dataset is stored in
dataset
print('Pima Indian Diabetes Dataset loaded...')
print('Total instances available :',len(dataset))
#length of dataset is 768 (number of rows)
print('Total attributes present :',len(dataset[0])-1)
#Attributes of dataset is 8 .
#Last column contains target value (diabetes-Yes/No)
27

28. for i in range(5):
print(i+1 , ':' , dataset[i])
# Print 1st
5 rows of dataset
28

29. 29
Pregnanci
es Glucose
Blood
pressure
Skin
thickne
ss
Insuli
n BMI
Diabetic
pedigre
e
function Age
Outcome(diabe
tes)
6 148 72 35 0 33.6 0.627 50 1
1 85 66 29 0 26.6 0.351 31 0
8 183 64 0 0 23.3 0.672 32 1
1 89 66 23 94 28.1 0.167 21 0
0 137 40 35 168 43.1 2.288 33 1
First 5 rows of dataset

30. splitRatio = 0.2
trainingSet, testSet = splitDataset(dataset, splitRatio)
#Function splitDataset split dataset into trainingSet
and testSet
# trainingSet contains 615 examples and testSet
contains 153 examples
print('nDataset is split into training and testing set.')
print('Training examples = {0} nTesting examples =
{1}'.format(len(trainingSet), len(testSet)))
# Display number of trainingSet examples and testSet
examples.
768 *0.2=153 testSet examples
615 trainingSet examples
30

31. • summaries = summarizeByClass(trainingSet)
summaries contains mean and standard deviation of
8 attributes of 2 classes.
Class 1
Class 0 mean.standard deviation
#prepare model
31

32. predictions = perform_classification(summaries, testSet)
153 instances
32
Using summaries (i.e model) testSet is predicted. i.e 153
examples are predicted using 615 trainingSet examples

33. • predictions = perform_classification(summaries,
testSet)
Outcome of each testSet(153) is predicted
33

34. • accuracy = getAccuracy(testSet, predictions)
• print('nAccuracy of the Naive Baysian Classifier is :',
accuracy)
# getAccuracy function will calculate the accuracy of
predictions of testSet
34

35. Training Set
35
8 examples
3 attributes

36. 36
Standard Deviation

37. 37
Standard Deviation

38. 38
Standard Deviation
Std Std Std

39. 39

40. 40
Std Std Std

41. 41