在for循环中使用数组python

Question

I am trying to run all the elements in just_test_data to all the elements in just_train_data , and return the lowest number, then run the new just_test_data through all the just_train_data , and so on until all the just_test_data has been run.

The error I keep getting is in the line

step_1 = (abs(just_test_data[i] - just_train_data[n]) ** 2)

IndexError: arrays used as indices must be of integer (or boolean) type

When I first try to run the loop.

import numpy as np
testing_data = np.genfromtxt("C:\Users\zkrumlinde\Desktop\Statistical Programming\Week 3\iris-testing-data.csv", delimiter= ',')
training_data = np.genfromtxt("C:\Users\zkrumlinde\Desktop\Statistical Programming\Week 3\iris-training-data.csv", delimiter= ',')

#create 4 arrays, the first two with the measurements of training and testing data
#the last two have the labels of each line
just_test_data = np.array(testing_data[:, 0:4])
just_train_data = np.array(training_data[:, 0:4])
testing_labels = np.array(testing_data[:, 4])
training_labels = np.array(training_data[:, 4])

n = 0
while n < len(just_train_data):
    for i in just_test_data:
        old_distance = 'inf'
        step_1 = (abs(just_test_data[i] - just_train_data[n]) ** 2)
        step_2 = sum(step_1)
        new_distance = np.sqrt(step_2)
        if new_distance < old_distance:
            old_distance = new_distance
            index = n
        n = n + 1
print(training_labels[index])

Answer 1

when you say for i in just_test_data: i will be the element itself, not the index.

you probably want something like for i in range(len(just_test_data)) this will have i as a number from 0 to the length of just_test_data - 1 .

edit: a few weird things in your code:

step_1 = (abs(just_test_data[i] - just_train_data[n]) ** 2)
step_2 = sum(step_1)
new_distance = np.sqrt(step_2)

this just returns abs(just_test_data[i] - just_train_data[n]) . are you meaning to add a ton of step_1 up and then eventually take the sqrt ? you need to check your indents.

old_distance = 'inf' is a string (pretty sure). you are probably looking for either np.inf or float('inf') . Also because you set this inside the for loop, it is getting reset for every i . you probably want it above 'for i in just_test_data:'

a quick pass at your code:

min_distance = np.inf
for n in range(len(just_train_data)):
    step_2 = 0
    for i in range(len(just_test_data)):
        step_1 = (just_test_data[i] - just_train_data[n]) ** 2
        step_2 += step_1
    distance = np.sqrt(step_2)
    if distance < min_distance:
        min_distance = distance
        index = n
print(training_labels[index])

This compares a point in just_train_data to all the points in just_test_data to compute a distance. It will print the minimum of these distances.

Answer 2

By using for i in just_test_data you're iterating through all the elements in the just_test_data array and not and index between 0 and the array length.

Also, it seems that your n = n + 1 line is not indented correctly.

Here's my guess for an updated version of your code:

import numpy as np
testing_data = np.genfromtxt("C:\Users\zkrumlinde\Desktop\Statistical Programming\Week 3\iris-testing-data.csv", delimiter= ',')
training_data = np.genfromtxt("C:\Users\zkrumlinde\Desktop\Statistical Programming\Week 3\iris-training-data.csv", delimiter= ',')

#create 4 arrays, the first two with the measurements of training and testing data
#the last two have the labels of each line
just_test_data = np.array(testing_data[:, 0:4])
just_train_data = np.array(training_data[:, 0:4])
testing_labels = np.array(testing_data[:, 4])
training_labels = np.array(training_data[:, 4])

n = 0
while n < len(just_train_data):
    for i in range(len(just_test_data)):
        old_distance = 'inf'
        step_1 = (abs(just_test_data[i] - just_train_data[n]) ** 2)
        step_2 = sum(step_1)
        new_distance = np.sqrt(step_2)
        if new_distance < old_distance:
            old_distance = new_distance
            index = n
    n = n + 1
print(training_labels[index])