Wine Quality Prediction

photo by Armands Brants on Unsplash

Importing the required modules

from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier 
from sklearn.metrics import accuracy_score
from sklearn.naive_bayes import GaussianNB
import matplotlib.pyplot as plt
from sklearn.svm import SVC
import pandas as pd
import numpy as np

Loading dataset

df = pd.read_csv(“datasets/winequality-red.csv”)

Understanding the dataset

df.head()

df.info()

Visualization

plt.bar(‘quality’,’fixed acidity’,data = df)

plt.bar(‘quality’,’volatile acidity’,data = df)

plt.bar('quality','citric acid',data = df)

plt.bar('quality','residual sugar',data = df)

plt.bar('quality','chlorides',data = df)

plt.bar('quality','free sulfur dioxide',data = df)

plt.bar('quality','total sulfur dioxide',data = df)

plt.scatter('quality','density',data = df)

plt.bar('quality','pH',data = df)

plt.bar('quality','sulphates',data = df)

plt.bar('quality','alcohol',data = df)

plt.plot(df["quality"])

Classifying the wine quality as good or bad

#bins will set the limits for the classification.
bins = (3, 6, 8) #qualities ranging from 3–6 are classified as bad and 6–8 as good
group_names = [‘bad’, ‘good’]
df[“quality”] = pd.cut(df[“quality”], bins = bins, labels = group_names)
df["quality"].head()

#one hot-encoding
df[“quality”] = pd.get_dummies(df[“quality”],drop_first=True)
df[“quality”][:5]

df[“quality”].value_counts()

Separating the dataset as target variable and feature variables

X = df.drop(“quality”, axis = 1)
y = df[‘quality’]
X.head()

y.head()

Splitting the dataset into train and test data

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 42)

converting y_train to 1-D

y_train = np.ravel(y_train)
y_train

Applying Standard scaling to get optimized result

sc = StandardScaler()
X_train = sc.fit_transform(X_train)
X_test = sc.fit_transform(X_test)

Creating the models

models = [ 'DecisionTreeClassifier', 'Support Vector Machine', 'GaussianNaiveBayes', 'KNeighborsClassifier', 'RandomForestClassifier']
accuracy_score_list = []

DecisionTreeClassifier

dtc = DecisionTreeClassifier()
dtc.fit(X_train, y_train)
pred_dtc = dtc.predict(X_test)
acc = accuracy_score(y_test, pred_dtc)
accuracy_score_list.append(acc)
print(acc)

Support Vector Machine

svm = SVC()
svm.fit(X_train, y_train)
pred_svm = svm.predict(X_test)
acc = accuracy_score(y_test, pred_svm)
accuracy_score_list.append(acc)
print(acc)

GaussianNaiveBayes

gnb = GaussianNB()
gnb.fit(X_train, y_train)
pred_gnb = gnb.predict(X_test)
acc = accuracy_score(y_test, pred_gnb)
accuracy_score_list.append(acc)
print(acc)

KNeighborsClassifier

knn = KNeighborsClassifier(n_neighbors=22)
knn.fit(X_train, y_train)
pred_knn = knn.predict(X_test)
acc = accuracy_score(y_test, pred_knn)
accuracy_score_list.append(acc)
print(acc)

RandomForestClassifier

rfc = RandomForestClassifier()
rfc.fit(X_train, y_train)
pred_rfc = rfc.predict(X_test)
acc = accuracy_score(y_test, pred_rfc)
accuracy_score_list.append(acc)
print(acc)

compare = pd.DataFrame({'Algorithms' : models , 'accuracy_score' : accuracy_score_list})
compare.sort_values(by='accuracy_score' ,ascending=False)

plt.plot(compare['Algorithms'], compare['accuracy_score'], label = "accuracy_score")

Maximum accuracy is 87% using Random Forest Classifier