Bachelor’s thesis

Information and Communications Technology

2020

Raj Kumar Tiwari

SENTIMENT ANALYSIS OF
ENGLISH TEXT USING
RASPBERRY PI
– Natural Language Text Processing, Raspberry
Pi, Sentiment Analysis
BACHELOR’S THESIS | ABSTRACT

TURKU UNIVERSITY OF APPLIED SCIENCES

Information and Communications Technology

2020 | 29

Raj Kumar Tiwari

SENTIMENT ANALYSIS OF ENGLISH TEXT USING

RASPBERRY PI
– Natural Language Text Processing, Raspberry Pi, Sentiment Analysis
IoT devices have been used in most of the areas in daily life these days. IoT devices are
the same devices used for single or multipurpose works in the home, offices, factories,
industries, machinery. The data collected from the operations of such devices are usually
collected and analyzed to generate further productivity and aid. Previously, the data were
analyzed by humans and calculators. Nowadays, the evolving technology and programs
help to understand data patterns and generate new algorithms which in turn provide
more sophisticated assistance with a more advanced approach.

An example of product development and its process is explained in this thesis to

understand the basic structure of IoT devices using natural language text processing to
create a basic understanding of the topic. For this purpose, an embedded device was
developed including a branch of machine learning to recognize the sentiment of English
Text.

The thesis aims to give a conceptual overview and explain the process itself. The
analysis of free text information comments and feedback helps to improve services and
understand people’s attitude towards the service with the help of a small processing unit
such as Raspberry Pi. The device use case is aimed at customer-oriented services and
it aims to collect the necessary feedback and analyze it. The device created is a basic
structure that can be further modified and developed using models such as face
recognition. The thesis explains the whole process of development and presents the
result.

KEYWORDS:

IoT Device, Machine Learning, embedded system, NLTP, text processing, Basic concept of
modern computing
CONTENTS

LIST OF ABBREVIATIONS

1 INTRODUCTION 6
1.1 Research Background 6
1.2 Research Methodology 7
1.3 Research objectives 7
1.4 Thesis statement 8
1.5 Thesis structure 8

2 CONCEPT OVERVIEW 9
2.1 Embedded System 9
2.2 IoT (Internet of things) 10
2.3 Network and Web Hosting 10
2.4 Natural Language Text Processing and Sentiment Analysis 11
2.5 Tools 12

3 PRODUCT DEVELOPMENT 14
3.1 Problem and Requirement 14
3.2 Environment set up and Installation 15
3.3 Development Methods and process 18
3.4 Database 20
3.5 Apache 24
3.6 Testing 25

4 DISCUSSION 26

5 CONCLUSION 27

REFERENCES 28
FIGURES

Figure 1. Raspberry PI set up with the interface ......................................................... 16

Figure 2. GUI of the Raspbian operating system ........................................................ 17
Figure 3. Display Label ............................................................................................... 19
Figure 4. Command-line interface to take input ........................................................... 19
Figure 5. Input Validation Error Message .................................................................... 20
Figure 6. Analyzer code that evaluates the input ........................................................ 20
Figure 7. Python string values to the database ........................................................... 20
Figure 8. Database storing the data from the python script ......................................... 21
Figure 9. MySQL Login ............................................................................................... 21
Figure 10. Database created for the first-time ............................................................. 22
Figure 11. Select the database and create a table and columns ................................. 23
Figure 12. Verify table creation ................................................................................... 23
Figure 13. Python data configuration .......................................................................... 24
Figure 14. PHP connection and data fetch method ..................................................... 24
Figure 15. File structure from the GitHub .................................................................... 25
Figure 16. Apache Web display .................................................................................. 25

TABLES

Table 1. Project Analysis ............................................................................................ 26

LIST OF ABBREVIATIONS

AI Artificial Intelligence

API Application Program Interface

Apt Advance Package Tool

ASIC Application-Specific Integrated Circuits

CPU Central Processing Unit

CSS Cascading Style Sheets

DSP Digital Signal Processors

FPGA Field-Programmable Gate Arrays

HTML Hyper-Text Markup Language

HTTP Hyper-text Transfer Protocol

I/O Input/ Output

IoT Internet of Things

MySQL My Structured Query Language

NLTP Natural Language Text Processing

PHP Hypertext Preprocessor

RAM Random Access Memory

RDBMS Relational Database Management System

SARP Sentiment Analyzer Raspberry Pi

SD Secure Digital
 6

1 INTRODUCTION

1.1 Research Background

Technology today has gone very far, advanced, with ease of access, and solved more
complex challenges with more efficiency, better and faster. With more exploration and
research, the growing sophisticated needs of humanity have driven the technology into
new realms such as AI, Embedded devices, web, and software programs.

Naturally, the development has become overwhelming at times for so many people to
adapt to new technologies and technical terms, features, works, new programming
languages, frameworks, and hardware devices. If we see it in terms of understanding
and using, more specifically adapting along with trends and technologies, such
adaptations have become complicated and confusing to consumers, users, academics,
geeks, students, and people from different areas.

The thesis aims to provide a necessary underlying concept of the IoT device and Natural
Language Text Processing (NLTP) by building a full-fledged example product. The
developed example product is a web server built upon Raspberry Pi and Raspbian Linux
distribution, which provides sentiment analysis of the text, based on NLTP as an
application. The developed product has been termed as SARP (Sentiment Analyzer on
Raspberry Pi) hereafter. The SARP is a byproduct of the process of investigating the
scope of an embedded device to run machine learning applications as web services.

With the advancement in technology, embedded devices are also having different inputs
of signals and data that have growing use by integrating such data with algorithms to
perform complex and distant tasks. The collected text data as inputs from the user or
machine can be helpful with natural language text analysis to create easier updates and
facilitate the organization or owner. A large amount of feedback and/or comments on a
service point and social media cannot be analyzed and calculated manually to take out
the needed output. In this place, the Raspberry Pi device with the NLTP model comes
into action to understand the sentiment pattern, calculate, and count the constant output
from the generated input pattern.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 7

The machines work smarter, faster, and more exact and help to automate the machines.
Text processing works to fulfill the bridge gap between the human processing of the text
to machine processing.

1.2 Research Methodology

This thesis is based on developmental research methods supported by knowledge

acquired from my previous academic learnings and technical experience of web
development, machine learning programming, and embedded systems. The theoretical
part consists of a description of tools and technology from different sources, as
mentioned in the references. The specific device with a similar application was not found;
that is why the thesis is explorative from the perspective of development. The data used
in the development are the random English sentences any subject speaks or the
comments and feedback from the internet pages for the testing purpose.

The Raspberry Pi was used because of its low cost, small size, and computing capability.
The software tools are chosen as per the compatibility of the need, task, and the
functionality of the product. The thesis work is primarily developmental research, which
aims to investigate the scope of IoT devices and NLTP for providing text processing
services as a web application. Developmental research is preferred research
methodology when researchers build a thing to address the given research problem
(Regmi, 2019). Therefore, developmental research methodology is used for this thesis
work. This thesis tries to address the research problem of investigating the scope of IoT
devices to implement a machine learning application by developing a prototype device.

1.3 Research objectives

This thesis's primary purpose is to provide a basic overview of the NLTP implemented in
an IoT device. Another objective is to help understand the concept, tools, and platforms
needed and used for the product development for the general readers and novice
students in the related field. The IoT device developed in this thesis lay a foundation of
a full-fledged IoT device that can be further developed with more features like image
processing, sound processing, text-processing for various purposes, and IoT related
possible topics. Precisely, the thesis also explains the terms - Embedded System, IoT,
Network and Web hosting, and Natural Language Text Processing.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 8

1.4 Thesis statement

Sentiment analysis is the interpretation and classification of emotions (positive, negative,

and neutral) within text data using text analysis techniques. The primary purpose of this
thesis is to use the Natural language text processing in Raspberry Pi and fetch the data
to the web, which can be further accessed via the Internet. In general, the reader of the
thesis will have a basic understanding of the IoT and an application of Machine learning.

1.5 Thesis structure

Chapter 1 introduces the thesis work and explains the key terms in an organized manner
used in the example product. Also, it establishes the connection between the intention
of the thesis writer and actual work. Chapter 2 describes concepts like Embedded
System, IoT, Network, and Case study.

Chapter 3 details the process of product development, such as the requirements

analysis, setting up and installing the environment to start the development work,
methods, and process of development and testing. Chapter 4 has documented the
discussion, whereas Chapter 5 concludes the overall thesis writing procedure, product
development, and results.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 9

2 CONCEPT OVERVIEW

2.1 Embedded System

General-purpose computers such as laptops and desktops are engineered to manage a

wide range of processing tasks, thus are not cost-efficient options to use where limited
processing power is required to perform specific tasks only. Embedded systems are
dedicated computers designed for a specific function or functions. Like general-purpose
computers, embedded systems are also a combination of hardware and software that
may be fixed in capability or programmable for a range of tasks. (What is an Embedded
System? - Definition from Techopedia, 2016)

The size, cost, power, consumption, reliability, and performance can be optimized by
engineers while designing.

Embedded system hardware

Embedded system hardware can be microprocessor- or microcontroller-based. The

integrated circuit plays a very significant role in real-time computing operations. (What is
an Embedded System? - Definition from Techopedia, 2016). The microcontroller that is
used here is Raspberry Pi. Microcontrollers consist of CPU, memory, flash memory,
RAM, and serial communication ports, whereas microprocessors constitute a central
processing unit. Microcontrollers can implement low powered full systems and are
frequently used on difficult and dedicated tasks.

Embedded System Software

Embedded systems are usually used as microcontrollers to carry out more

straightforward computations within less memory-intensive program environments.
Embedded systems often use operating systems with basic functionalities such as
Rasbian for Raspberry Pi or tailored programming language platforms such as Arduino
IDE for Arduino and custom minimal operating system. Generally, programs and
operating systems are stored either on flash memory or rewritable flash memory. (What
is an Embedded System? - Definition from Techopedia, 2016). The Raspberry Pi has
used the Raspbian operating system, which is a Linux distribution. Raspbian is a Debian
Linux distribution ported to Raspberry Pi. On top of that, the task design to carry out the
necessary input and analysis is written on python, PHP, and web framework languages.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 10

2.2 IoT (Internet of things)

The network of embedded devices and other microcontrollers, when connected to the
Internet, is known as the Internet of Things (IoT). Here the things not only include IoT
devices such as embedded systems but also IoT enabled physical assets such as
vehicles with GPS trackers. (The Internet of Things (IoT) - essential IoT business guide,
2020).

IoT refers to the devices connected to the web, but the definition has been changing
over time. Nowadays, IoT refers to the smart devices that communicate with each other,
collecting and sharing data across the Internet. IoT devices provide an additional level
of intelligence to the device handling and processes without human intervention every
time. For Example, A refrigerator can check the stock of the food and order it to the
grocery store. (What is embedded system? - Definition from WhatIs.com, 2020) (Ranger,
2020)

2.3 Network and Web Hosting

Computer networks and the World Wide Web are the central infrastructures of
information and communications technology (ICT). The term 'web hosting' is used to
indicate the process of deploying web content on the Internet. It includes maintaining a
web server to store web content and managing the configuration of the necessary
technology to make the stored content available via the Internet. Some of the necessary
configurations are registering a domain name and associating it with a unique IP address
so that the web content can be accessible in a web address. (What Is Website Hosting?,
2020)

Web server stores data in any form and serves data in the form of (HTML) web pages
for client browsers. In this project, we have deployed a website using the localhost
network provided by the local router. Apache webserver hosts the files within the same
network in this project.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 11

2.4 Natural Language Text Processing and Sentiment Analysis

NLP is a branch of artificial intelligence that studies interactions between computers and
human languages. Among various applications of artificial intelligence, NLP deals with
applications that require computers to process human languages. Such applications
include speech recognition, content summarization, content translation, spam detection.
(Yordanov, 2018)

The smartphone with speech recognition uses NLP to understand what is being said.
Also, many people use laptops whose operating system has built-in speech recognition.

Sentiment analysis, a new field in Natural Language Processing, aims at identifying the
intent of the content, and classifying opinions and sentiments. “Sentiment analysis
studies people's sentiments, opinions, attitudes, evaluations, appraisals, and emotions
towards services, products, individuals, organizations, issues, topics, events, and
attributes. The text is classified on two different bases, such as polarity of the sentiment
and polarity of the outcome."(D'Andrea, Ferri, Grifoni and Guzzo, 2015).

Different classification approaches and tools are used for sentiment analysis implemen-
tation.The Sentiment analysis is a process involving several steps such as Data Collec-
tion, Text preparation, Sentiment detection, Sentiment classification and displaying the
result. Data collection from different user generating points such as content from blogs,
forums, social networks, and customer-oriented service points. This raw data collected
from primary sources are not organized, unstructured or in a complex language like
slangs, streets, communities varying in contextual meaning. The manual analysis of the
data is not a very good idea or is considered a traditional way. So, the raw data is to be
cleaned and prepared so that the machine understands it easily. Therefore, text analytics
and natural language text processing are used to extract and classify. Non-textual, irrel-
evant and stop words are eliminated from the cleaning process. The rest of the data are
examined on a subjective or objective basis. Subjective sentences are classified in pos-
itive, negative, good, bad; likes, dislikes, but classification can be made by using multiple
points. Sentiment analysis turns unstructured text into meaningful information. After the
analysis, the test results are displayed on tables, percentages, wording (Happy, Neutral
and Sad), graphs like a pie chart, bar chart and line graphs. (D'Andrea, Ferri, Grifoni and
Guzzo, 2015).

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 12

2.5 Tools

Python

Python language is a high-level programming language with extensive use in a different

context of modern computing. Python is an object-oriented, interpreted language with
dynamic semantics. It is high-level built-in structures, dynamic typing, and binding makes
the application development process easier and faster. Python reduces the cost of
program maintenance thanks to its simplicity and easy to learn syntax. Python also
supports modules and packages. It is open source and freely distributed. (What is
Python? Executive Summary, 2020).

Due to its features and efficiency, the python programming language has been used
extensively in machine learning. Several developers have contributed to its huge support
of libraries and functions. The analyzer that has been used in this project uses python
language.

Vader Sentiment Analyzer

"VADER (Valence Aware Dictionary and Sentiment Reasoner) is a lexicon and rule-
based sentiment analysis tool specifically attuned to sentiments expressed in social
media." (vaderSentiment, 2020). VADER uses a combination of lexical features that
generally analyze the words with their semantic orientation as Happy and sad.

It has produced impressive outputs analyzing texts from reviews, comments, social
media, product reviews, and newspaper editorials. It is available in python language that
is used in this project. Vader can identify the sentiment, describes how polarized the
sentiment is, and provides the score. (Staff, 2020).

PyMySQL Connector

PyMySQL is a library used to interact with the database library. It is a driver required to
connect the python language script to the database.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 13

MYSQL Server

MySQL is an open-source relational database management system (RDBMS) that helps

to separate and store data in tables for software applications. MySQL database uses
Structure query language to communicate with the database and has a client-server
architecture. Databases are stored in files that provide optimal speed. MySQL database
server is fast, easy, scalable, reliable, and works simultaneously in computers alongside
other applications. (Regmi, 2019).

MariaDB is installed for creating, storing, and serving the data from and to the database.

Apache

Apache is an open-source and widely used web server software. Apache is fast, reliable,
secure, free, and runs on 67% of all web servers in the world. Extensions and modules
are used to highly customize the program to meet the needs of services running of
different environments. (Pandey, 2018).

PHP

PHP (Pre-HyperText Processor) is installed to handle the server-side. PHP is a server-

side scripting language that allows generating web pages dynamically and
programmatically by allowing one to embed code within HTML templates (Regmi, 2019).
The PHP script connects to the database and displays the input and Results in a
webpage.

HTML and CSS

HTML and CSS are used to design the Graphical User Interface. HTML creates the page
template design and frame. CSS provides the customized and required styles to the
template page. PHP processes the data and generates HTML pages.

Rufus

Rufus is a tool that helps to write the memory stick with iso image files of operating
systems. Rufus is used for writing the file system image containing the operating system
and other applications to the SD card in this project.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 14

3 PRODUCT DEVELOPMENT

3.1 Problem and Requirement

Research problem: The research problem is to investigate the scope of an embedded

device to run machine learning applications as web services.

Research goal: The goal is to develop a prototype product implementing NLTP on

Raspberry Pi.

Use case scenario:

Machine learning applications on embedded devices provide cheaper alternatives for

enterprises and business to use such services in a cost-efficient way. The importance of
machine learning implementation in factories, industries, and corporations to identify
correct person, flawless product classification pushing the computation to the edge has
been advancing and widespread. The sentiment analysis of text helps to understand the
social phenomenon and bring more clarity of human communication and understanding.
This kind of device can be helpful in monitoring classes with teachers, students,
hospitals, and shopping mall feedback machines and can be modified with more
sophisticated machine learning algorithms. The device software requires English text as
input to analyze its sentiments.

Requirements,

Hardware: stand-alone IoT device. Chosen Raspberry Pi.

Software: OS, application to run a web server, application to perform sentiment analysis
of the text.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 15

3.2 Environment set up and Installation

Setup

The necessary hardware such as Raspberry Pi, screen, keyboard, mouse, and HDMI
Cable was collected, to begin with, the setup process. The 16GB memory SD card from
the Raspberry Pi is written with a popular Raspbian operating system using Rufus. Rufus
is a popular memory writing software in a wide variety of memory sticks (MySQL ::
MySQL 8.0 Reference Manual :: 1.3.1 What is MySQL?, 2020).

After the writing process is done, the SD card is inserted into the Raspberry Pi slot. The
Raspberry Pi is connected to the screen via HDMI cable. The keyboard and mouse are
connected to the Raspberry Pi USB slots. The power is connected to monitor and
Raspberry Pi. Then the Raspbian operating system is shown on the monitor. The first
part of the set up ends here.

Figure 1. Raspberry PI set up with the interface

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 16

Installation

The Raspbian operating system provides the GUI and command line support to install
software like Apache, MySQL, Python, and other required libraries. We prepare the
development environment after installing the necessary tools mentioned above.

Figure 2. GUI of the Raspbian operating system

Steps:

The terminal can be found on the upper-left corner forth in figure 2.

Inside the Terminal.

Command to insert:

Apache:

sudo apt install apache2 -y

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 17

Python:

pip install MySQL-python

Vader Sentiment Analyzer:

sudo pip3 install vaderSentiment

PyMySql Connector:

sudo apt-get -y install python-mysql.connector

MariaDB Server:

sudo apt-get install MariaDB-server

PHP

sudo apt-get install PHP

After the installation, the development environment is ready to code and develop
software.

Graphical User Interface

The GUI is designed to display the result that is stored in the database. HTML and CSS
are used to design the front end. The GUI has two display labels, each of which displays
the user input and the analyzed result.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 18

Figure 3. Display Label

3.3 Development Methods and process

The development of SARP consist of python and PHP script, and one style files. All the
program files are stored inside the var/www/HTML folder. The script file “rest.py” does
the task of taking input from the user.

Figure 4. Command-line interface to take input

The input is validated to check if it is empty or number. It gives the error message
("Nothing or number has been entered. Please try again").

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 19

Figure 5. Input Validation Error Message

If the input passes the validation, the input is assigned to variable user_string. The
user_string is then passed as a parameter to analyze the polarity of the value. The Vader
sentiment analyzer checks the polarity and intensity of the user_string and stores the
result in the score variable.

Figure 6. Analyzer code that evaluates the input

The result is then measured on the 100 scales. The result is identified as Negative, Sad,
and Happy or Positive derived from the resulting scale. The result is then saved to the
database.

Figure 7. Python string values to the database

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 20

Figure 8. Database storing the data from the python script

The configuration method is discussed in the Database configuration chapter.

3.4 Database

The database called raspberry_pi_data with a table name nltk_resullts is created with
two columns. The User_input and Results are the two columns inside the table.

The prerequisite installation for MySQL is completed in the installation section. The login
process is mandatory to get into the MySQL Command line.

The following command is used in the terminal.

$ sudo mysql -p root -u

It required a password that we set during the installation. After the password is entered,
the MySQL command line is accessed.

Figure 9. MySQL Login

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 21

Here, the database is created, accessed, and other operations can be performed.

The basic MySQL command is to show databases to check the existing database. This
command gives the result of the existing databases.

> show databases;

Figure 10. Database created for the first- time

In Figure 10, the result list is missing raspberry_pi_data. In the next line, the database is
created with the following command.

> CREATE DATABASE raspberry_pi_data;

As Figure 10 displays, the database has been created.

The database needs tables and columns to store different kinds of data the program is
going to need. This project requires a single table with two columns.

Firstly, it is required to select the database and execute the further operation of creating
tables.

Select the database

> use raspberry_pi_data;

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 22

And create the tables:

Figure 11. Select the database and create a table and columns

The id is int type, UserInput and Results are the varchar type. These are the kind of data
types that are used in programming to separate the nature of data.

The following command can verify the table is created or not.

> select * from the nltk_results;

Figure 12. Verify table creation

The result is an empty set because there is nothing in the database. The data comes
after the teaching of the language python and PHP configures with the database.

Database Configuration

The PHP script and python script use the same database raspberry_pi_data and both
languages have their way to connect to the database.

Python connection method:

Figure 13. Python data configuration

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 23

Python uses a pymysql connector to connect with the database. The mariadb.connect
function has the parameter host, user, password, database, and values assign to it. The
same variable is executed with the cursor function. It is the python configuration method.

After the database is connected, several SQL operations such as creating, reading,
putting, updating, and deleting can be performed with the database.

Php connection method:

Figure 14. PHP connection and data fetch method

PHP does the same task of configuration with the above code. The database credentials
are assigned to the PHP variable, which then is again passed as a parameter to PDO
Object inside the try-catch function. The connection method in PHP tries to connect and
return Connection failed error in case of a fail database connection.

After the database is connected, the database can be accessed, and operations such as
creating, reading, delete, putting, and updating can be performed from the code.

The PHP code finally fetches the data from the database with the help of the SELECT
statement in Figure 4, from connection and fetch method.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 24

Below is the file structure from Github.

Figure 15. File structure from the GitHub

This is a single-page interface with a PHP code that configures the connection between
the database and brings the data to the user in the web interface.

3.5 Apache

Finally, All the files are stored inside the var/www/html directory. The apache is started
with the following command.

$ sudo service apache2 restart

After the apache is started, the page is available to be hosted on localhost. In this
project, the localhost address is 192.168.1.172. The index.php is added at the end of
the address to retrieve the file.

Figure 16. Apache Web display

The page is ready to serve the results.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 25

3.6 Testing

The testing is done, and the SARP as hardware is fully functional to this task. The
software has been tested with seventy-five number of statements in two datasets. The
seventy-five number of datasets vary in sentiment and randomly chosen. The test results
are shown below in the table.

Table 1. Project Analysis

Number of Datasets 75
Positive 48
Negative 3
Neutral 24
Accuracy 96.5%

From the Result analysis table, the analyzer has proven to be 96.5% accurate to identify
the input within the test data. The critical note to always keep in mind is that the result is
relative to this test data and can be different with different varieties of data.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 26

4 DISCUSSION

Results

The result of the product development is the device name "SARP," which takes the
English text as input and displays the sentiment of the input to the user and on the web.
The SARP takes the user input and categorizes the input into three types of sentiment
and output the compound results.

The three types of sentiment are Negative, Neutral, and Positive. These results are
calculated based on the polarity and intensity of each word in the input. The result
produced by the SARP has an accuracy of 96.5% with the testing dataset.

Use Case and Benefits

The SARP is a kind of feedback collecting and analyzing device where users can input
text, paragraph, and files of English literature and get the result as Happy, Negative, and
Normal. This device can be used in hospitals to get customer feedback where it does
the task of storing and displaying the results to necessary service owners.

"The World Wide Web is an immense collection of linguistic information that has in the
last decade gathered attention as a valuable resource for tasks such as machine
translation, opinion mining and trend detection, that is, Web as Corpus." (Kilgarriff and
Grefenstette, 2003). It can be used to analyze such data from the web, in Markets, Malls,
Hotels, shops, cafeteria to gather feedbacks for the construction and monitoring of the
service. Moreover, the module is the prepared underlying architecture that can be used
with different other modules. The SARP helps people live and work smarter and get a
better understanding between people, service customers, and create better results. The
SARP provides companies' systems work, delivering insights into everything from
individual performance to users.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 27

5 CONCLUSION

In this chapter, we will discuss the conclusion based on our research and development.
The purpose of the thesis was to clarify the concept and development process via
theoretical research and product development. The goal was achieved through different
intermediate processes of research, literature review, tutorials from different sources,
and practical implementation. The final output was the product which can take any
English statement as input text, process it via natural language text processing module,
sends it to the internet, and can be accessed via a web user interface. During the process
of writing and development, the fundamental structure was a full-fledged device. The
Command-line works as the user input interface. When the script rest.py is run, then the
user can input any text or paragraph. After the input is submitted, the script analyzes the
given input sentiment and store the results in the database. The result is displayed on a
webpage.

This thesis introduced an embedded device module that can analyze the sentiment. It
introduces tools such as Rufus, Raspbian operating system, PHP, Python, MySQL,
Apache web server, and the methods to use them. The SARP successfully demonstrates
that the product can take input and display the result on a web page. The thesis also
provides a theoretical description of the tools and technologies used for product
development.

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 28

REFERENCES

A. Mishra, D. Patil, N. Karkhanis, V. Gaikar and K. Wani, "Real time emotion detection from
speech using Raspberry Pi 3," 2017 International Conference on Wireless Communications,
Signal Processing and Networking (WiSPNET), Chennai, 2017, pp. 2300-2303, doi:
10.1109/WiSPNET.2017.8300170.

D'Andrea, A., Ferri, F., Grifoni, P. and Guzzo, T., 2015. Approaches, Tools and Applications for
Sentiment Analysis Implementation. International Journal of Computer Applications, 125(3),
pp.26-33.

Dev.mysql.com. 2020. Mysql :: Mysql 8.0 Reference Manual :: 1.3.1 What Is Mysql?. [online]
Available at: <https://dev.mysql.com/doc/refman/8.0/en/what-is-mysql.html> [Accessed 18 April
2020].

GraziCNU, M., 2017. [online] Turning Your Raspberry Pi Into a Personal Web Server. Available
at: <https://www.instructables.com/> [Accessed 20 March 2020].

IoT Agenda. 2020. What Is Embedded System? - Definition From Whatis.Com. [online] Available
at: <https://internetofthingsagenda.techtarget.com/definition/embedded-system> [Accessed 25
February 2020].

i-SCOOP. 2020. The Internet Of Things (Iot) - Essential Iot Business Guide. [online] Available at:
<https://www.i-scoop.eu/internet-of-things-guide/> [Accessed 15 March 2020].

Kilgarriff, Adam & Grefenstette, Gregory. (2003). Introduction to the Special Issue on the Web as
Corpus. Computational Linguistics. 29. 333-347. 10.1162/089120103322711569.

Networksolutions.com. 2020. What Is Website Hosting?. [online] Available at:

<http://www.networksolutions.com/education/what-is-web-site-hosting/> [Accessed 1 March
2020].

Pandey, P., 2018. Simplifying Sentiment Analysis Using VADER In Python (On Social Media
Text). [online] Available at: <http://Medium.com> [Accessed 10 March 2020].

PyPI. 2020. Vadersentiment. [online] Available at: <https://pypi.org/project/vaderSentiment/>

[Accessed 20 May 2020].

Python.org. 2020. What Is Python? Executive Summary. [online] Available at:

<https://www.python.org/doc/essays/blurb/> [Accessed 5 March 2020].

Ranger, S., 2020. What Is The Iot? Everything You Need To Know About The Internet Of Things
Right Now | Zdnet. [online] ZDNet. Available at: <https://www.zdnet.com/article/what-is-the-
internet-of-things-everything-you-need-to-know-about-the-iot-right-now/> [Accessed 27 February
2020].

Staff, A., 2020. What Is Apache? - What Is A Web Server?. [online] WPBeginner. Available at:
<https://www.wpbeginner.com/glossary/apache/3> [Accessed 7 March 2020].

Techopedia.com. 2016. What Is An Embedded System? - Definition From Techopedia. [online]

Available at: <https://www.techopedia.com/definition/3636/embedded-system> [Accessed 7
February 2020].

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari

 29

Yordanov, V., 2018. Introduction To Natural Language Processing For Text. [online] Available at:
<http://Medium.com> [Accessed 1 March 2020].

TURKU UNIVERSITY OF APPLIED SCIENCES THESIS | Raj Kumar Tiwari