H. J.

THIM TRUST’S

THEEM COLLEGE OF ENGINEERING, BOISAR

DEPARTMENT OF COMPUTER ENGINEERING

2022-23

TE 6TH (CE)

ASSIGNMENT NO 1: CASE STUDY

SUBJECT: ARTIFICIAL INTELLIGENCE

SUBJECT TEACHER: SHAHEGUL AFROZ

CASE STUDY ON

“FRIEND RECOMMENDING SYSTEM ON SOCIAL MEDIA

APPS WITH ARTIFICIAL INTELLIGENCE WAY”
SUBMITTED BY:

1.Mohammad Tanzeel (27)

2.Atul Tiwari (61)
1. Introduction
When the technology was not a boom twenty years ago people typically made friends with them who
live or work close to themselves such as neighbors and colleagues. Such friends are termed as the G-
friends that is physical position base friends, the name is given because the friends were made on the
basic of the physical distances between each other. With the progress in the social networks like
Facebook, twitter, Google+ and many more they have supplied with radical ways of making friends.
As a statistic provided by Facebook on a standard every user has 130 friends which is larger than the
old concept

To suggest a regular friend to the user is the major dispute to the presented scheme. In the conventional
scheme the technique depends on the pre-offered user associations to select friend candidates. For
example, Facebook depends on a social connection analysis among those users who already share
common friends and recommends equal users as potential friends. Unluckily, this advance might not
be the suitable based on most modern study. For recommending friends there are a small number of
conventions which are behavior, routine, attitudes, tastes, ethical principles, economic level and
people they already know. this Rules are widely used by all the recommendation scheme and can offer
with the top outcome but it is not easy to collect behavior or life style of a person on the things he
does on the net. If there is technique that could collect information on user’s every day routines and
actions, then that can be broken rules and suggest friends to people based on their related life styles.
The recommendation structure can be deployed as a separate application on Smartphone or to existed
frameworks. Then this method can grasp the daily life and then the recommendation can be specified
easily.

Life style consists of day-to-day activities of human, there may be hundreds of actions which can
shape a significant succession and aid the life style system to grasp. The word activity is used to
specially submit to the actions taken in the order of seconds, such as “sitting”, “walking”, or “typing”,
while the expression life style is used to specify higher-level concepts of daily lives, such as “office
work” or “shopping”. For instance, the “shopping” life style mostly consists of the “walking” activity,
but may also contain the “standing” or the “sitting” tricks.

The figure shows an example model. It describes the analogy between people’s daily lives and
documents. Here daily life can be treated as the document life style as topics and activities as the
words. So, if any other user has the same or near to same lifestyle is recommended.

In this system the user’s comments likes and post profile are analyzed as these three somewhat related
to the life style of the user as these are also activities which users do most on the social networks.
Once this method calculates the similarity percentage this system will be able to recommend friends
to the users based on the similar match or the most likely to be a friend
2. Existing System

Till now all the recommendation systems works on two basic principles one is the user likeness and
mutual friends and the other one is the community liked system. The recommendation of the Facebook
is on the basis on the mutual friends. More the mutual friends the more users are recommended as the
friends. But this may or may not be the case as in real life it is not important that having too many
mutual friends can make the two of them friends too. Similar recommendation can be seen in the
movie recommender too.

Here the recommendation is based on the group of people. For example, if a particular movie is liked
by 70 people from 100 than the user who still not have seen the movie are recommended for the movie
which is mostly loved. But this system too fails as the taste and likes of each and every user is different,
they may not like the movie or the movie is then just a bad recommendation from the system.

Most of the pre-existing relationship to pick up the friends. Facebook relies on social link analysis
among those who already share common friends and recommends symmetrical user as common
friends Existing social network system recommends friend to user based on the social graph which
may not be the most appropriate to reflect user’s preference on friend selection on real life.

Recommendation systems that try to suggest items (e.g., music, movie, and books) to users have
become more and more popular in recent years. For instance, Amazon recommends items to a user
based on items the user previously visited, and items that other users are looking at. Netflix and Rotten
Tomatoes recommend movies to a user based on the user’s previous ratings and watching habits.
Recently, with the advance of social networking systems, friend recommendation has received a lot
of attention. Generally speaking, existing friend recommendation in social networking systems, e.g.,
Facebook, LinkedIn and Twitter, recommend friends users if, according to their social relations, they
share common friends.

2.1 Problems with Existing System

 One challenge with existing social networking services is how to recommend a good friend to a
user. Most of them rely on pre-existing user relationships to pick friend candidates.
 Facebook relies on a social link analysis among those who already share common friends and
recommends symmetrical users as potential friends.
 Unfortunately, this approach may not be the most appropriate based on recent sociology findings .
3. Proposed System

A friend recommendation system for social networks, which recommends friends to users based on
their life styles instead of social graphs. Friend recommendation system discovers profile and daily
activities of users, measures the similarity of life styles between users, and recommends friends to
users if their life styles have high similarity. We model a user’s daily life as life documents, from
which his/her life styles are extracted by using the Jaccard distance algorithm. Similarity metric to
measure users,
and calculate users Impact in terms of life styles with a friend-matching graph. We integrate a linear
feedback mechanism that exploits the user’s feedback to improve recommendation accuracy. There
are still multiple challenges for extracting users’ life styles and recommending potential friends based
on their similarities. First, how to automatically and accurately discover life styles from database?
Second, how to measure the similarity of users in terms of life styles? Third, who should be
recommended to the user among all the friend candidates? To address these challenges, in this paper,
we present Recommending Friend on Social Network with Artificial Intelligence.

The contributions of this work are summarized as follows:

1) To the best of our knowledge RFSNAI is the first friend recommendation system exploiting a user’s
life style information discovered from database?
2) Inspired by achievements in the field of text mining, we model the daily lives of users as life
documents and use the probabilistic topic model to extract life style information of users.
3) We propose a unique similarity metric to characterize the similarity of users in terms of life styles
and then construct a friend-matching graph to recommend friends to users based on their profile and
daily activities.
4) We conduct both small-scale experiments and large-scale simulations to evaluate the performance
of our system. Experimental results demonstrate the effectiveness of our system.

3.1 Overview

Our society is becoming more and more dependent on social networking sites, which nowadays are
used in everyday life, from business to banking, from entertainment to health care. In this section, we
give a high-level overview of the Friend book system.

On the client side, each user can record data of its user, perform real-time activity recognition and
report the generated life documents to the servers. It is worth noting that an offline data collection and
training phase is needed to build an appropriate activity classifier for real-time activity recognition.
We choose MySQL as our low-level data storage platform and Hadoop Map Reduce as our
computation infrastructure. On the server side, seven modules are designed to accomplish the task of
friend recommendation. The data gathering module collects life documents from user’s system. The
life styles of users are extracted by the life style analysis module with the probabilistic topic model.
Then the life style indexing module puts the life styles of users into the database in the format of (life-
style, user) instead of (user, life-style). A friend-matching graph can be constructed accordingly by
the friend-matching graph construction module to represent the similarity relationship between users’
life styles and their different characteristics of a family based on system.
4. Future Work
In this paper we presented the design and implementation of recommending friends on social
network with artificial intelligence approach. It’s totally different from the friend recommendation
mechanisms relying on social graphs in existing social networking services, this system extracted
life styles which generally includes profile post and comments of user and recommended potential
friends to users if they share similar life styles. We implemented recommendation system and
evaluated its performance on both small-scale experiments and large-scale simulations. The results
showed that the recommendations accurately reflect the preferences of users in choosing friends.

5. Conclusion

This work explores the potential of network representation in the field of friend recommendation
and points to several directions in the future. First, the attention mechanism will be used to enhance
the user feature learning in the graph convolution layer, so as to improve the performance of friend
recommendation. The second is to comprehensively process the user’s friend deletion information,
take into account the negative information and time factors, use the reinforcement learning content
plus the user’s decision layer, and make decisions on whether he will make friends with others or
will delete a friend. Third, it can also be combined with explain able recommendation system to
strengthen the interpretably of the deep learning model of friend recommendation. Better user
representation is expected to get to complete the tasks recommended by friends efficiently.

6. References

[1] Sundong Kim, “Friend Recommendation with a Target User in Social Networking Services
“In proceeding of IEEE International Conference on Data Engineering (ICDE), Feb2012, PP.
235-239

[2] Jeff Naruchitparames, Mehmet Hadi Guand Sushil J. Louis “Friend Recommendations in
Social Networks using Genetic Algorithms and Network Topology “in proceeding of CEC,2011,
PP.29-42.

[3] Tanvi Shruti” Social Recommender systems: Improving recommendations through

personalization” Under the direction of Professor Steven Lindell Haverford College. Computer
Science Department. Spring 2011

[4] Pooja Tasgave, Amit Dravid. “Friendbook: A Lifestyle based Friend Recommendation System”
in proceeding of International Journal for Research in Applied Science & Engineering Technology
(IJRASET) March2015, PP. 560-564