Wine Quality Prediction using Machine Learning

Project report submitted in partial fulfilment of the requirement

for the degree of Bachelor of Technology

in

Computer Science and Engineering/Information Technology

By

Aryan Negi (171220)

Parvesh Sharma (171256)

Under the

supervision of

Dr. Hari Singh Rawat

to

Department of Computer Science & Engineering and Information

Technology
Jaypee University of Information Technology Waknaghat,
Solan- 173234, Himachal Pradesh
 Table of Content
 Declaration by Candidate……………………...…..I
 Acknowledgement…………………………………II
 Certificate by Supervisor…………………………III
 Abstract……………………………………………IV
 List of Abbreviations………………………………V
 List of Figures……………………………………..VI
 Chapter 1…………………………………….……VII
 Chapter 2…………………………………………VIII
 Chapter 3…………………………………………..X
 Chapter 4…………………………………………XV
 Conclusion……………………………………...XVII
 Reference…………………………………….....XVIII
 Candidate’s Declaration
I hereby declare that the work presented in this report entitled “Wine Quality
Prediction using Machine Learning” in partial fulfilment of the requirements for
the award of the degree of Bachelor of Technology in Computer Science and
Engineering/Information Technology, submitted in the department of Computer
Science & Engineering and Information Technology, Jaypee University of
Information Technology Waknaghat is an authentic record of my own work
carried out over a period from July 2020 to December 2020 under the supervision
of Dr Hari Singh Rawat, Assistant Professor, Computer Science & Engineering
and Information Technology.
The matter embodied in the report has not been submitted for the award of any other
degree or diploma.

Aryan Negi(171220)

Parvesh Sharma(171256)

This is to certify that the above statement made by the candidate is true to the best
of my knowledge.

Dr. Hari Singh Rawat

Assistant Professor
Computer Science & Engineering and Information
Technology Dated:
 ACKNOWLEDGMENT
We would foremost extend our gratitude towards our supervisor Dr
Hari Singh Rawat at the Department of Computer Science &
Engineering and Information Technology at Jaypee University of
Information Technology, under whom this project has been
conducted. We would like to thank him for the help he has been
giving throughout this work.

We have grown both academically and personally from this

experience and are very grateful for having had the opportunity to
conduct this study.

We are also thankful to all other faculty members for their constant
motivation and helping us bring in improvements in the project.

Finally, we like to thank our family and friends for their constant
support. Without their contribution it would have been impossible to
complete our work.

Aryan Negi, Parvesh Sharma

JUIT Waknaghat
July 2020

II

I
 Certificate

This is to certify that Aryan Negi (171220) and Parvesh Sharma

(171256) have completed their Minor Project in Machine Learning
under the guidance of Dr. Hari Singh.

Dr. Hari Singh

Assistant Professor(Senior Grade)
CSE and IT
Dated:

III
 ABSTRACT
The main goal of this project is to predict wine quality whether it is good or
bad. For centuries tasting has been done by humans and they have always
predicted on the basis of sensory organs. But in recent times the industries are
adopting newer technologies and applying them in all kinds of areas. But, still
there are many areas in which human expertise is needed like product quality
assurance. Nowadays, it becomes an expensive process as the demand of
product is growing over the time. Therefore, this project searches different
machine learning techniques such as MLP classifier, Decision Tree classifier,
Support Vector Machines (SVM) for product quality assurance. These
techniques do quality assurance process with the help of available
characteristics of product and automate the process by minimizing human
interference.

IV
 List of Abbreviations

ACRONYM DEFINITIONS

1. SVM Support Vector Machines

2. KNN K-Neighbour Nearest
3. MLP Multi-Layer Perceptron
4. SGD Stochastic Gradient Descent

V
 List of Figure

S.No Table of content

1. Figure 1: Histogram

2. Figure 2: Boxplot

3. Figure 3: Scatter Plot

4. Figure 4: Heat Map

VI
 Chapter – 1

1.1 INTRODUCTION
The most defining period of human history will always be remembered as computing
moved from mainframes to PCs to cloud and now to artificial intelligence. An
important area of artificial intelligence which came in lime light, called as Machine
Learning, allows computers to get into some kind of self-learning mode involuntary.
With the concepts and ideas from machine learning, we have been able to spread from
miscellaneous accurate reduplications to big data iteration that too with at a
marvellous speed. This spectacle has been in momentum over the last several years.
On the other hand, data mining includes data discovery and sorting it among large
data sets vacant to identify the required designs and begin affiliations with the aim of
answering teething worries over and done with data analysis. Basically linking, device
learning and data mining use the same type of method and set of processes, except the
kind of data pre-dealing out and end guess varies. Between these two core expanses to
predict and present the truest results potential.

1.2 PROBLEM STATEMENT

Predicting on the test data of Red Wine Quality Dataset and finding the accuracy of
the model using Logistic Regression, involving import of dataset, quality check on the
data (Data Wrangling), and performing Exploratory Data Analysis (Univariate and
Bivariate Analysis) using Histograms, Boxplots and Scatter Plots. Thus, modelling
the dataset using various machine learning algorithms.

1.3 OBJECTIVE

 Build a Jupyter notebook in Anaconda, import data, and view numbers loaded
obsessed by the notebook.
 Practice Pandas to clean and formulate data.
 Use scikit-learn to create the machine learning exemplary.
 Use Matplotlib to see the model's performance.

VII
 Chapter -2

Jupyter notebooks are highly collaborative, and since they can take in
executable enigma, they provide the seamless platform for manipulating data
and edifice predictive models from it.

1. Firstly we download the dataset from the Kaggle.

2. In the notebook's following cell, enter the following Python code to

load winequality-red.csv, craft a Pandas DataFrame from it, and ceremony
the first five commotions.

3.Connect the Run button to execute the code. Sanction that the output
remind us of the output below.

VIII
The Data Frame that we formed contains information of all the contents
percentages that are present in red wine and the wine quality as well. It has
more than 1000 rows and 12 columns. (The output says "5 rows" because
Numbers Frame’s head job only returns the first five rows.) Each row embodies
the amount of content available in the wine as well as it’s quality as well .We'll
mine at the data more closely a bit later in this segment.

IX
 Chapter-3

A Dataframe is a two-dimensional characterized data structure. The

columns in a Files Frame can be of changed types, just like columns in a
binge sheet or catalogue table. It is the most commonly used object in
Pandas. In this exercise, we will observe the Data Frame and the data
inside it more thoroughly.

1. One of the first possessions you archetypally want to know about a

dataset is how many dins it contains. To get a calculation, type the
resulting statement into an bare cell at the end of the notebook and run
it:

2. Yield a flash to survey the 12 columns in the dataset. Here is a ample list
of the columns in the dataset.

Column Explanation

Fixed acidity Percentage of Fixed acidity in wine

Volatile acidity Percentage of Volatile acidity in wine

Citric acid Percentage citric acid in wine

Residual sugar Percentage of residual sugar in wine

chlorides Percentage of chlorides in wine

Free sulphur dioxide Percentage of Free sulphur dioxide in wine

Total sulphur dioxide Percentage of Total sulphur dioxide in wine

X
 Column Explanation

density Percentage of Density in wine

pH Percentage of pH in wine

sulphates Percentage of sulphates in wine

alcohol Percentage of alcohol in wine

quality Quality of Wine

The dataset takes in an even dispersal of quantities of various substances used in

making of a particular wine and it’s quality. The substances used in the dataset
are often commonly measured in making of a particular wine and after the wine
has been made it’s quality is checked and accordingly scored.

One of the most central aspects of fixing a dataset for practise in apparatus
learning is decide on the "feature" columns that are significant to the outcome
we are trying to predict while filtering out columns that do not affect the
outcome, could bias it in a negative way, or might produce multicollinearity.
Another important task is to exclude missing values, either by accordingly
scoring them or by filling them with the average value of that column. In this
exercise, we will check for missing value rows/columns.

1. One of the first things data scientists typically look for in a dataset is
missing values. There's an easy way to check for missing values in Pandas.
To demonstrate, execute the following code in a cell at the end of the
notebook:

XI
 2. The next step is to find out where the missing values are. To do so,
execute the following code:

Confirm that we see the following output listing no count of missing values in
each column:

The dataset is now "clean" in the sense that missing values have been replaced and
the list of columns has been narrowed to those most relevant to the model.

3. Univariate and Bivariate Analysis is carried out on the dataset to discover

patterns. The below figures show univariate analysis using histogram and
boxplot for ‘fixed acidity’ and ‘quality’ columns, respectively.

Fig1:-Histogram for one variable

XII
12
 Fig2: -Box plot for one variable

The below figures show bivariate analysis using scatter plot for ‘fixed acidity’
and ‘quality’ columns which reflects uniformly spread data.

Fig3: -Scatter plot for two variables

XIII
A heat map is extremely powerful way to visualize relationships between
variables in high dimensional space. For example, in this case a correlation
matrix with heat map colouring is shown below. A correlation matrix is a
table showing correlation coefficients between sets of variables. Each
random variable in the table is correlated with each of the other values in the
table. This allows us to see which pairs have the highest correlation.

Fig4: - Heat Map

XIV
 Chapter-4

To fashion a machine learning model, we want two datasets: one for training
and one for testing. In practice, we often have only one dataset, so we split it
into two. In this exercise, we will perform an 70-30 split on the dataframe we
prepared in the previous lab such that it can used to train a machine learning
model. We will also isolate the dataframe obsessed by feature columns and
label supports. The former has the columns used as input to the wine(for
example, the fixed acidity, alcohol, sulphates, etc. ), while the latter contains the
stake that the model will try to predict — in this case, the quality column, which
indicates whether a flight will arrive on time.
1. In a new cell at the end of the notebook, enter and execute the following
statements:

The first statement imports scikit-learn's train_test_split helper function. The

second line uses the function to split the DataFrame into a physical activity
set having 70% of the unusual data, and a test set enclosing the left over
30%.

2. Now use this command to show the numeral of rows and columns in the
DataFrame comprehending the feature columns used for tough:

XV
There are various types of machine learning copies. One of the most public is
the KNN model, stores all the available data and classifies a new data point
based on the similarity. This means when new data appears then it can be easily
classified into a well suite category by using K- NN algorithm.

One of the doles of using scikit-learn is that we don't have to build these
reproductions — or implement the algorithms that they use — by hand. Scikit-
learn includes a variety of classes for instigating collective machine learning
models. One of them is Decision Tree classifier, which organizes a series of test
questions and conditions in a tree structure.

1.Execute the following code in a new cell to create a Decision tree

classifier object and train it by calling the fit method.

2.Now call the predict method to test the model using the values in X_test,
followed by the score method to determine the mean accuracy of the model:

The accuracy is 99%, which seems good on the surface.

In the real world, a trained data scientist would look for ways to make the model
even more accurate. Among other things, they would try different algorithms
and take steps to tune the chosen algorithm to find the optimum combination of
parameters. Another likely step would be to expand the dataset to millions of
rows rather than a few. But for our purposes, the model is fine as-is.

XVI
18
20
 CONCLUSION

Machine Learning is a technique of training machines to perform the activities a

human brain can do, albeit bit faster and better than an average human-being.
Machine Learning can review large volumes of data and discover specific
trends and patterns that would not be apparent to humans. For instance, for an e-
commerce website like Amazon, it serves to understand the browsing behaviors
and purchase histories of its users to help cater to the right products, deals, and
reminders relevant to them. It uses the results to reveal relevant advertisements
to them. Machine Learning algorithms are good at handling data that are multi-
dimensional and multi-variety, and they can do this in dynamic or uncertain
environments.

Machine Learning requires massive data sets to train on, and these should be
inclusive/unbiased, and of good quality. There can also be times where they
must wait for new data to be generated. ML needs enough time to let the
algorithms learn and develop enough to fulfil their purpose with a considerable
amount of accuracy and relevancy. It also needs massive resources to function.
This can mean additional requirements of computer power for you. Another
major challenge is the ability to accurately interpret results generated by the
algorithms. You must also carefully choose the algorithms for your purpose.

Machine Learning can be incredibly powerful when used in the right ways and
in the right places

XVII
 REFERENCE

Dataset has been taken from following link:

https://www.kaggle.com/uciml/red-wine-quality-cortez-et-al-2009

Research papers:

Wine Quality Prediction using Machine Learning

Algorithms by Devika Pawar, Aakansha Mahajan,
Sachin Bhoithe

Links:

1.https://www.verzeo.in/
2.https://www.tutorialspoint.com/machine_learning/wh
at_is_machine_learning.htm
3.https://towardsdatascience.com/exploratory-data-
analysis-8fc1cb20fd15
4.https://en.wikipedia.org/wiki/Machine_learning

XVIII