KNN Implementation for predicting Diabetes Outcome
Classification using KNN
Objective
The objective is to predict whether a patient has diabetes,based on diagnostic measurements.
This dataset is originally from the National Institute of Diabetes and Digestive and Kidney
Diseases. It includes diabetes outcome and measurements of Glucose, Insulin, BMI et al."I will
apply kNN machine learning on presented measurements to predict the diabetes outcome. The
data download link https://www.kaggle.com/mathchi/diabetes-data-set
KNN algorithmis a supervised machine learning algorithm that deals with similarity. KNN
stands for K-Nearest Neighbors. The number K is typically chosen as the square root of N, the
total number of points in the training data set.
KNN has the advantage of being nonparametric. It’s basically a classification algorithm that
will make a prediction of a class of a target variable based on a defined number of nearest
neighbors.
Data Descriptions
Using k-nearest neighbors to predict the Diabete Outcome
•STEP 1: Load the Libraries
•STEP 2: Load the data set
•STEP 3: Exploring and preparing the data
– Examine the structure of the Dmdata frame
– from the results we notice the unusual scenario of zero in Glucose, BP,
SkinTHickness, Insulin and BMI; presumably those are NA
– Replace thoses zero into NA
– Examine the missing value (the percentage of missing)
– From the results, we noticed more than 40% Insulin missing, those missing will
be imputed with the value of the nearested neighbors via knnImpute
– the structure of data
•
