Shulou(Shulou.com)06/02 Report--

How to achieve a KNN algorithm, many novices are not very clear about this, in order to help you solve this problem, the following editor will explain in detail for you, people with this need can come to learn, I hope you can gain something.

Teach you how to implement the KNN algorithm yourself

KNN algorithm, also known as K-neighbor algorithm, can solve regression and classification problems, but solving classification problems is its advantage.

The essence of KNN algorithm is to find k samples that are similar to the data we provide, then judge the labels of these k samples, and finally count the number of times each tag appears, and the tag with the largest number of times will be regarded as the label of the data we provide.

Let's start with the workflow:

Machine learning is based on data, so it is necessary to convert the physical object into the form of vector, matrix or tensor.

The distance between the test sample and other samples is calculated by Euclidean distance.

Sort the distance from small to large, and take the first K values

Judge the corresponding labels of the first K values, and make statistics.

The label with the most statistics is the prediction result.

Now let's make it happen.

Import all libraries or modules that need to be imported first

# Import the dataset from sklearn import datasets# import counter included with sklearn, which is used to count the number of label occurrences. From collections import Counter# is used to split the dataset from sklearn.model_selection import train_test_split# and to calculate the Euclidean distance import numpy as np.

Second, import the iris dataset and store sample features and tags

Data = datasets.load_iris () # Store sample features to XX = data ['data'] # Store sample tags to YY = data [' target']

Then, in order to evaluate the model later, the data set is divided into training set and test set.

X-ray recording, testing, writing, writing, testing, writing, etc., etc., etc. = train_test_split (Xmagedrine Yreign recording statekeeper 2000)

Set random_state to a fixed value, so that the result of each run will be the same, which helps us to judge the bug that occurs in the process.

Third, write a function eus_dis that calculates the distance

Def eus_dis (instance1, instance2):''calculate the distance between two samples instance1: Array instance2: Array distance = np.sqrt (sum ((instance1-instance2) * * 2)) return distance

Then, really start to implement the KNN algorithm.

Def KnnClassify (XGravity Y training for x in X k):''implement KNN algorithm X: feature training set-- > array type Y: label training set-> array type test: feature test set-> array type k: adjacent value-- > int''# calculate the distance between samples distances = [eus_dis (x, test) for x in X] # sort from small to large And take the first K values. Return subscript kneighbors = np.argsort (distances) [: K] # count the number of occurrences of each tag count = Counter (Y [kits]) # return the tag return count.most_common () [0] [0] that has the most occurrence

Finally, let's test it and observe its accuracy.

# Storage the prediction results of the feature test set by the model predirect= [KnnClassify (X-ray prediction pr ecnology Yoshihizhong 5) for test in X_test] # calculate the number of prediction results equal to the actual results count = np.count_nonzero ((predirect==Y_test) = = True) print ("the prediction accuracy of the model is:% .3f"% (count/len (X_test). Is it helpful for you to read the above content? If you want to know more about the relevant knowledge or read more related articles, please follow the industry information channel, thank you for your support.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.