Introduction to K-NearestNeighbors classifier:
K-nearest neighbor classifier is one of the introductory supervised classifier, it addresses the pattern recognition problems, and also the best choices for addressing some of the classification related tasks.
K-nearest neighbor classifier algorithm was proposed by Fix & Hodges in 1951 for performing pattern classification task.
K-NearestNeighbours Pseudocode:
KNN code using Python:
Euclidean distance:
The euclidean distance is presented as follow, and in my code I will import numpy for basic math functions:
KNN class:
The following is the K-NearestNeighbors class with three methods.
Test the built KNN algorithm on Spotify dataset:
After building the algorithm, I will use Spotify dataset from Kaggle. The goal is to predict whether or not I would like a song, using my KNN then compare the score with built in Sklearn KNN.
Upload the dataset:
You can Download dataset from here
Print the first 5 rows to check how our dataset looks:
Split the dataset to train and test sets:
Run the test with both algorithms:
Before testing the algorithms, the training and testing sets will need to be converted to numpy arrays
Results using the built KNN:
Results using sklearn KNN:
Conclusion:
In conclusion, this was a step by step building K-Nearest Neighbors algorithm by using base python. After testing and comparing the Sklearn KNN and the built algorithm, the same score was given. Although, there is a quite large difference in the run time between the two tests.