Machine-Learning-with-Python/knn_iris_scratch.py at master · sssalam1/Machine-Learning-with-Python · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
## KNN scratch code on iris Data

# Importing the libraries
import pandas as pd
import numpy as np
import statistics
# from sklearn import preprocessing
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

# Import Iris Data
data = pd.read_csv("file:///E:/py_practice/Iris.csv")
data = data.iloc[0:100,:]
x_data = data.drop(["Species","Id"],axis=1)
y_data = data.Species.map({"Iris-setosa":0,"Iris-versicolor":1})

# Train Test Split
x_train, x_test, y_train, y_test = train_test_split(x_data,y_data,test_size=0.2)

x_train = x_train.values
x_test = x_test.values

# Calculating all the distances for traning example
k = (1,3,5)
total_acc = []
for v in k:
    clas = []
    for i in range(len(x_test)):
        d = []
        for j in range(len(x_train)):
            d.append(np.sqrt(sum((x_train[j,:]-x_test[i,:])**2)))
        clas.append(statistics.mode(y_train.values[np.argsort(d)[range(v)]]))

    accuracy = accuracy_score(y_test,clas)
    total_acc.append(accuracy*100)
#print(total_acc)

best_k = k[np.argsort(total_acc)[len(total_acc)-1]]
print("Best k is ", best_k)