Question

We observed 10 samples of the products. The column A-C are the attributes of the 10 products, and the column D is the category of the products.

Now we have a new product, with the following stats:

Length: 1.5

PctPos: 0.5

PctNeg: 0.5

Please use KNN to classify this new prodcut. Please use K = 3 (3 nearest neighbors).

Length	PctPos	PctNeg	Category
1.66	0.84	0.28	C
1.62	0.86	0.25	A
2.04	0.7	0.23	B
1.68	0.78	0.03	A
1.62	0.76	0.95	A
1.68	0.88	0.02	A
2.08	0.66	0.29	B
1.82	0.66	0.58	C
2.06	0.68	0.53	B
1.8	0.64	0.41	C

please provide codes

Answer 1

import csv
def loadDataset(filename, trainingSet=[]):          #file reader function
   with open(filename, 'rt') as csvfile:
        lines = csv.reader(csvfile)
        dataset = list(lines)
        for x in range(len(dataset)-1):
            for y in range(3):
                dataset[x][y] = float(dataset[x][y])
            trainingSet.append(dataset[x])

trainingSet=[]
testSet=[]
print ('Train: ' + repr(len(trainingSet)))
print ('Test: ' + repr(len(testSet)))

def euclideanDistance(instance1, instance2, length):
   distance = 0
   for x in range(length):
       distance += pow((instance1[x] - instance2[x]), 2)
   return pow(distance,0.5)

data1 = [1.66, 1.62, 2.04, 'a']     #put all the given length in data1 list.
data2 = [0.84,0.86,0.7, 'b']         ##put all the given pctpos in data2 list.
distance = euclideanDistance(data1, data2, 3)
print ('Distance: ' + repr(distance))

def getNeighbors(trainingSet, testInstance, k):
   distances = []
   length = len(testInstance)-1
   for x in range(len(trainingSet)):
       dist = euclideanDistance(testInstance, trainingSet[x], length)
       distances.append((trainingSet[x], dist))
   distances.sort(key=lambda x:x[1])
   neighbors = []
   for x in range(k):
       neighbors.append(distances[x][0])
   return neighbors

trainSet = [[2.2, 2.4, 2.1, 'a'], [4.9, 4.1, 4.3, 'b']]
testInstance = [5, 5, 5]
k = 1                                                     #testing of code
neighbors = getNeighbors(trainSet, testInstance, 1)
#print(neighbors)

def getResponse(neighbors):
   classVotes = {}
   for x in range(len(neighbors)):
       response = neighbors[x][-1]
       if response in classVotes:
           classVotes[response] += 1
       else:
           classVotes[response] = 1
   sortedVotes = sorted(classVotes.items(), key=lambda x:x[1], reverse=True)
   return sortedVotes[0][0]

neighbors = [[1.2,1.8,2.2,'a'], [2.3,2.9,2.4,'a'], [3.1,3.4,3.5,'b']]
response = getResponse(neighbors)

#predictions = ['a', 'a', 'a']

def main():
    trainingSet=[]
    testSet=[[1.75,72,35],[1.82,82,42]]
    loadDataset('sample.data', trainingSet)      #create a sample.data test file.
    print ('Train set: ' + repr(len(trainingSet)))
    print ('Test set: ' + repr(len(testSet)))
    predictions=[]
    k = 3
    for x in range(len(testSet)):
        neighbors = getNeighbors(trainingSet, testSet[x], k)
        print(neighbors)
        result = getResponse(neighbors)
        predictions.append(result)
        print('> predicted=' + repr(result))
    accuracy = getAccuracy(testSet, predictions)

main()    #by calling the main function the program starts running and for better understanding open the code

              #in jupyter notebook.