Option #1 k-Nearest Neighbor Suppose you work for a wine company as a data mining expert. You are to build a k-nearest neighbor algorithm to predict the quality of wine from different factors....

Wine Quality Data Analysis
The objective of this report is to analyse read wine samples, build a kNN classification algorithm and evaluate the model on the dataset. The inputs include objective tests (e.g. PH values) and the output is based on sensory data (median of at least 3 evaluations made by wine experts). Each expert graded the wine quality between 0 (very bad) and 10 (very excellent). Following attributes are available in the dataset:
Wine Quality Dataset Independent Variable
● Fixed acidity
● Volatile acidity
● Citric acid
● Residual sugar
● Chlorides
● Free sulfur dioxide
● Total sulphur dioxide
● Density
● pH
● Sulphates
● Alcohol
Dependant Variable
● Quality
Data Pre-processing
We replace the Quality variable with a binary class variable for the classification task according to the following rule:
Class: High when Quality > 5
Class: Low when Quality <= 5
The summary statistics of all the variables in the dataset, we notice that a few of the features are on a different scale compared to the rest. Therefore, during in the model evaluation section, we test our kNN model on data that has been normalized to make the model less sensitive to the scale of certain features.
The following table shows us the Highly correlated pairs.
Data Transformations:
Outlier Analysis: We will check the outlier present in our dataset. We remove those records from our dataset where at least one attribute’s value for that record is an outlier. Since some outlier convey valuable information to model, the threshold that we set to decide an outlier is if the value is greater than 3 standard deviations from the mean of the attribute.
Data Normalization: We notice that a few of the features are on a different scale compared to the rest. Therefore, we need to normalize the data to make the model...