Chatbot as a Personal Assistant

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 20 (2018) pp. 14644-14649

14644

Gayatri Nair

Department of Computer Science and Engineering, SRM Institute of Science and Technology

(Deemed to University u/s 3 of UGC Act, 1956), Chennai, India.

Soumya Johnson

Department of Computer Science and Engineering, SRM Institute of Science and Technology

(Deemed to University u/s 3 of UGC Act, 1956), Chennai, India.

V. Sathya (Guide)

Assistant Professor, Department of Computer Science and Engineering,

SRM Institute of Science and Technology (Deemed to University u/s 3 of UGC Act, 1956), Chennai, India.

Abstract

Chatbots usually work as an optimiser of customer service.

Chatbots are designed in such a way that the users are made to

believe that they are talking to an actual human being but

rather they are talking to a machine. The main advantage of

this explicit characteristic of a chatbot is that it can be given a

virtual personality of its own just like a specific person of a

particular profession. This paper proposes an idea of an

artificially intelligent chatbot as a personal assistant that can

serve as your resume. Also, this chatbot helps the user in

fixing meetings and adding them in the calendar.

Our chatbot

is basically a resumeBot. It is a resume in the form of a

chatbot. . A resume in the form of chatbot is the new

generation in the field of job applications. A chatbot will help

the recruiter to know your personality along with all your

qualifications and personal details. The chatbot will

communicate with the user with the help of natural language

processing and AIML files.

Keywords: Artificial Intelligence, Chatbot, Personal

Assistance.

INTRODUCTION

A chatbot also known as a talkbot, IM bot, chatterbot, Bot,

interactive agent, or an Artificial Conversational Entity.

Chatbot is a computer program that helps you to communicate

via messages. They are formulated to mimic human behavior.

They can reproduce exactly how a human would act as a

conversational partner.[3]

A chatbot as a personal assistant will be able to not only

control your schedule on Google calendars but also will be

able build an immaculate resume.

A personal assistant controls and handles all the home

appliances with a single touch. Chatbots are those computer

programs that mimic human conversation using artificial

intelligence and are transforming the way people interact with

the Internet for information and assistance and knowledge. It’s

not surprising that chatbot apps are considered the new

browsers and the new websites. Google Now Cortana, Siri and

Alexa are all examples of chatbots. And Facebook has

opened up Messenger, making it not merely an app but a

platform upon which developers can build an entire chatbot

system. Facebook Messenger provides over the 30,000

chatbots and it also stays as their main habitat.

In this today’s generation, it is really important to stand out

from the crowd and make a mark. It is really important to

prove your worth in this era of competition. Having a basic is

just not enough. So, the best way to make yourself stand out

of the crowd is to use latest technology. By using the

technology like Artificial Intelligence will surely make you

and your professional profile to stand out. So presenting your

resume in the form of chatbot is the best way of implementing

Artificial Intelligence to make your professional profile better.

However, chatbots have a big potential for implementation in

personalized corporate accounts on every social media

platforms. Chatbots are designed to be the ultimate virtual

assistant, helping you to accomplish various tasks ranging

from answering questions, getting driving directions, turning

up the thermostat in your smart home, or play your favorite

tunes and even draft a question and answer survey.

Chatbots are being made to ease the pain that the industry is

facing today. The purpose of chatbots is to support and help to

scale business teams in their relationship with customers and

consumers at the same time.

OVERVIEW

Existing System

Additionally, there are some existing systems from which the

idea of creating an application of chatbot (virtual personal

assistant) was inspired.

1. A.L.I.C.E. One of the most famous chatbot which works on

Pattern Matching Strategy is the Artificial Linguistic Internet

Computer Entity (A.L.I.C.E.) . The AIML files for A.L.I.C.E.

are available online which contain categories like music, art,

philosophy, etc. So for the basic working of our chatbot, these

AIML files are being used. Also, another original AIML file

for the category “Meetings” has been generated, which

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 20 (2018) pp. 14644-14649

14645

answers specific meeting related questions. Thus, as the

project concentrates more towards the scheduling module,

AIML files are being used for the pattern matching framework

of chatbot.[1]

Proposed System

We have made a chatbot that will change the way you look at

a resume. Our chatbot is basically a resumeBot. It is a resume

in the form of a chatbot.

In past years we have not seen much changes in LinkedIn

profiles. Our LinkedIn profiles surely give a good impression

and are more interactive than resumes but they are not

interactive in a way a chatbot is. A resume is normally a list of

dates and details of the candidate written down. But what if

the resume you build is actually interactive. A resume in the

form of chatbot is the new generation in the field of job

applications. A chatbot will help the recruiter to know your

personality and actually see examples of your work. Chatbots

give a sense of personality of a person instead of just a word

document with some facts written down.

SYSTEM

ARCHITECTURE

There are four modules in the architecture.

Presentation layer

ii.

Service layer

iii.

Data access layer

iv.

Database layer

Presentation layer:

In this interface the user communicates with the

chatbot. The output generated in this layer of the

system is served as the input of the next layer i.e.

service layer.

ii.

Service layer:

The service layer consists of Gmail API and the

calendar API and various other APIs.

iii.

Data access layer:

This is the intermediary layer of the system. The

function of this layer is to exchange the data between

the database and the chatbot. It uses pattern matching

algorithm to select mosr favourable response and

sends it to the presentation layer.

iv.

Database layer:

It comprises of three main databases. The first

database is of the AIML files i.e. the AIML database.

The second is the information stotred about the user

that includes all the information in his resume from

name to hobbies. The third database is the user’s

Google calendar.

DESIGN

CONCEPT

The design of a Chatbot is represented here:

Figure 1:

Use Case Diagram of Chatbot Design.

A Chatbot can be designed by these steps:[2]

Selection of OS

We have Windows OS since it’s the most user friendly and

best OS in the market till date.

Selection of Software:

Eclipse software is used for programming in java because of

its great user friendly nature. It is compatible with multiple

platforms beimg a java IDE.

A. Creating a Chatbot:

Chatbot is created by writing a program in java language.

Chatbot can also be created without any involvement of

coding for beginners. Websites like Chatfuel, Botsify and

Moton.ai helps one to create their own chatbot without any

coding.

B. Creating a Chat:

A chat is created between the user and the chatbot. The chat is

easily understandable by the bot.

C. Pattern Matching

The whole concept of a chatbot lies on its ability to match

various given patterns. It is the speciality of artificial

intelligence which is used during the designing of a chatbot.

During pattern matching the input of the text which is given

by the user is compared with all the texts stored in its

database. After executing the comparison recognition an

output is generated accordingly.

D. Conversational and Entertaining

The chatbot responds in a friendly way and that too in English

language. If you ask about any data regarding the user the

chatbot will answer it almost instantly.And that too in the best

and most accurate way possible. It is just as talking to the

USER

CHATBOT

Ask

questions

Receive

responses

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 20 (2018) pp. 14644-14649

14646

same person whose resume is loaded in the database and

commanding a person to check the calendars.

 Creation Platform

Creation platforms provide a wide range of software

foundations, frameworks and toolkits like APIs. There are

other features like, natural language processing, searching,

and image processing. These platforms can be distribution-

platform-specific or can be used across multiple platforms,

such as the Microsoft Bot Framework, Botkit, and

Pandorabots. These provide services from documentation and

code templates to even no-code-required bot-building and

managing interfaces just as Chatfuel. Chatfuel is one of the

aamazing feature that lets one create chatbot apps without any

advanced coding and is very helpful in creating prototypes.[3]

 Distribution Platform

Distribution platforms tells us about where and how users will

access bots. These days many are centered around messaging

or social networking like Messenger, Skype, and

WeChat,Telegram etc.. Other platforms which are domain

specific mainly concentrate on developers for example, Slack,

Teams, and HipChat.

These platforms support human–bot, bot–bot , or even

system–bot interactions.These platforms define and improvise

on how users interact with bots and have built-in support for

commands, natural language, speech, and more rich and

efficient UI controls. The interaction with a bot strongly

influenced by the experience of the user and the types of tasks

users can perform with chatbot. Some of the third-party sites

like BotList and ChatBot provide online catalogs of bots. It

makes easier for developers to promote and market their bots.

Some distribution channels have monetization features that let

bots safely collect payments from users, which is particularly

useful for people developing transactional style bots.[3]

The new areas of technology can be unleashed by making this

into reality.

IMPLEMENTATION

The chatbot is an exceptionally made app that can be used as a

personal assistant. When a person logs into the given app it

asks you for some personal details regarding one’s name,

alarm time and mainly the to do list.

You can also send mails and messages automatically if the

app is connected with your email-id. This helps get a new

arena of the technology into perspective.

Chatbot is an application of the computer which uses artificial

intelligence to imitate human conversation. It helps the user

by answering the questions asked by them. The program is

implemented using Java programming language. Particularly,

one uses the Java applets. Applets are used as it is easy for it

to create dialog boxes needed for communication between the

user and the chatbot. Detailed implementation is given below :

A. Fundamental Design Techniques and Approaches

Creating the dialog box

The variety of packages needed for making the dialog box are

transfered. The text area and size of dialog box is already

given. To scroll the screen during the conversation a vertical

scrollbar is used. As the size of the horizontal dialog box is

not fixed it is not used. The conversation always takes place

vertically and never horizontally.

B. Creating a database

Two dimensional string arrays are used to build a database to

store data. Arrays contain rows and columns. All the rows are

used for request and response. Even rows contain request and

all the odd rows contain answers or responses. Columns in the

given array are applied to store and save various types of

questions that a chatbot has to answer. Default responses are

stored in a different row in the array. And this row is not used

for storing any other data. When the matching question is not

found in the array then the default response statements are

used.

C. Modules Description

The description of the modules are as follows[6]:

 Chatbot()

In this function, all the variables that are used for creating the

dialog box is added.

There is a default close operation set to EXIT_ON_CLOSE so

that the dialog box closes whenever it is exited. It requires

background colour. And it is set using the built set

Background() function.

 Random()

The input is given by the user. and the same is accepted using

text(). By using trim() function all the punctuations are

removed. Uppercase letters are converted to lowercase. There

is a variable response. And this is used to hold a byte value.

This value is set to 0. When the response is 0, the database

finds a match for input given. It is then returned as a response

and is later displayed on the dialog box in the text area. And if

the response is 1 then no match is found in the database. In

such cases a default response will be returned. The function

Random() is used chose responses stored in database[7].

 AddText()

This function is used to add all the strings and texts used for

input and output. They are added in text area part of the dialog

box.

 InArray()

This is specifically used to match the pattern. Variable

matches are used to hold the Boolean values. It is then set to

false. If a match is found for inputs given by the user in the

database then true is returned orelse false is returned. The

result is now returned to keypressed() function. This result

will then be displayed on the dialog box.

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 20 (2018) pp. 14644-14649

14647

 Google APIs:

Google APIs are program interfaces. They were developed by

Google. They help us in authorization with services provided

by Google. These APIs include gmail, maps, search etc. Third

party apps can use these APIs for authorization. If they need

to access Google’s services [9].

 Calendar API:

Anyone can use Google calendar APIs to find out and view

calendar events. Both public and private events are accessed.

Our main function is to set up meetings and updates like a to-

do list. Also our chatbot can be set so the user is reminded of

his schedule regularly. Events on this calendar can also be

acquired and altered easily at the click of a button.[10].

EQUATIONS

AND

ALGORITHM

Chatbots have always fascinated humans with there self

operating technology. Day by day chatbot technology is

becoming more like humans and more automated along with

being more interactive. Chatbots are known for their

impulsive responses and persistent connectivity. These

tempting features have made them capable of replacing the

ongoing trend in web application.[5]

Working of a chatbot:

The input that the user gives in the form of text is first

analysed and processed by a “classifier”. A classifier is a

software function present in the system. The output of the

classifier is then associated with a conversational intent. This

intent is the one that is responsible for the responses of a

chatbot.

Figure 2. Example of a chatbot.

A classifier can be thought of a method of categorizing a

sentence under some categories. The sentences are the input

from the user and the categories are the intent which is

responsible for a response. So, if a user asks “How are you”

then it can be considered as a classifier and responses like

“I’m good” or “All fine” can be considered as intents.

Classifiers can be divided into three types.

i. Pattern matchers

ii. Algorithms

iii. Artificial Neural network

i. Pattern Matchers

Pattern matching is the most effective and commonly used

concept in chatbots. Pattern matching is used to categorize a

text input given by the user to generate appropriate response

for the user. AIML stands for “Artificial Intelligence Markup

Language”. It can be considered as typical framework for

pattern matching.

An easy pattern matching example:

Figure 3. Example for pattern matching

ii. Algorithms

In chatbot technology for each kind of input given by the user

a distinct pattern is stored in the database. These patterns help

in providing users a suitable response. The use of algorithms

is to reduce the classifiers and to create a large combination of

patterns. These patterns generally follow a hierarchy which

creates an effective and manageable structure. These

algorithms follow the basic algoritm called Multinational

Naive Bayes.

For example, assume a set of words are given which belong to

a specific class. With each new input of a word, each word is

counted for its occurrence. And each class is assigned a score.

The highest score in the class is the closest to be linked with

the input word.

Few sample Input sentence classification:

Input - “Hello good morning”

term - “hello” (no matches)

Term - “good” (class: greeting)

term: “morning” (class: greeting)

classification: greeting (score=2)

With the help of such an equation, word matches are noted for

each class. Classification scores identify the class with the

most term matches. But it also has certain limitations. The

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 20 (2018) pp. 14644-14649

14648

score explains that which intents to most likely match the

sentence. There is no certainity that it is perfect.

iii. Artificial Neural Networks

Artificial Neural Network are used to calculate the output

from the input given by the user with the help of weighted

connections. These weighted connections are generated from

recurrent iterations during the processing or training of the

data. At each step of training or processing the data the

weights are modified. These modification in weights result in

an output with utter accuracy.[4]

Figure 4. Artificial neural network.

In Artificial Neural Network, each sentence given as input is

divided into separate words. Each word is then used as an

input for the neural network. There can be many different

kinds of neural network and pattern matchers but as the

complexity increases the output of the chatbot becomes more

accurate.

NLU (Natural Language Understanding):

It consists of three concepts:

 Entities:

Entity can be considered as the main ideology of your chatbot.

It can be a payment system chatbot or customer support

chatbot or a resume bot as proposed in this paper.

 Intents:

Intents are basically the outputs or the responses that a

chatbnot gives on encountering an input by the user. In short

these are the actions that a chatbot performs when a user gives

a text input.

 Context:

In NLU algorithm whenever a sentence in processed or

scanned, the system does not have any information about the

past conversations of the user. it means if a user asks a user,

and the chatbot has given a reponse to that question, but, the

chatbot will not have a record of the question that has been

just asked. So for distinguishing the parts of the conversation,

the states of the conversation are recorded. A state can be a

flag like “school studied” or parameter such as “educational

qualification”.

NLP (Natural Language Processing)

In Natural Language processing (NLP) chatbot uses a

collection of steps that can be a collection of questions to

transform the user’s input text into relevant data and then give

an appropriate answer or response.

These are the steps involved in Natural Language Processing:

 Sentiment Analysis:

This analysis keeps a track on the conversations to check if

the user is having a good quality experience.

 Tokenization:

The NLP separates the sentence into different words or

tokens. NLP classifies and divides a string of data into small

parts or tokens.

They are differently used for application purposes.

 Normalization:

In this step the chatbot looks for misspelled words or any kind

of typing mistakes in the text input by the user.

 Dependency Parsing:

In this step the chatbot finds out the subject, verb and object

of the sentence given by the user. It looks for nouns and

related or dependent phrases in the user’s text input to

understand what the user is trying to convey.

Similar to all the web applications present today, chatbots also

need to have a database. The database is the collection of

information and facts that is used to generate a suitable and

most appropriate response to the user. the dat related to all the

activities are stored in the database. So, NLP converts

common human language to relevant information. The

information can be a collection of patterns and relevant texts

that can be applied to get a proper response.

CONCLUSION

Therefore, a chatbot can be used to not only to chat and get

information but it takes artificial intelligence in a whole new

light. The main advantage of a chatbot is that it can provide

services at any given rate. They respond immediately to the

users’ demands with precisely relevant information.

This enhances the rate of the communication operation. In

simple terms, chatbots make your service faster and as a

personal assistant it is the best. You can use it for the

betterment of common people.. They can also be used to

instantly book or buy tickets, rooms, tables and also set

alarms, notes, send reminders regarding meetings, alarms for

medication timings and various other reasons.

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 20 (2018) pp. 14644-14649

14649

REFERENCES

[1] Namita Mhatre, Karan Motani, Maitri Shah, Swati

Mali, “Donna Interactive Chat-bot acting as a

Personal Assistant” in International Journal of

Computer Applications (0975 – 8887) Volume 140 –

No.10, April 2016.

[2] M. Dahiya, “A Tool of Conversation: Chatbot” in

International Journal of Computer Sciences and

Engineering, Volume-5, Issue-5.

[3] Carlene Lebeuf, Margaret-Anne Storey, and Alexey

Zagalsky, “Software Bots” in IEEE Software,

www.computer.org/software.

[4] “ How Do Chatbots Work? A Guide to the Chatbot

Architecture”, https://www.marutitech.com/chatbots-

work-guide-chatbot-architecture/.

[5] Gk_, “Soul of the Machine: How Chatbots Work.”,

https://medium.com/@gk_/how-chat-bots-work-

dfff656a35e2.