CONTENT

 Introduction
 Artificial Intelligence in
Recommendation System
 Examples of Recommendation
System
 Benefits of AI in
Recommendation System
 Common Challenges
 Future Work
 Conclusion
 References
 Introduction
The explosive growth in the amount of available digital information and the number
of visitors to the Internet have created a potential challenge of information overload
which hinders timely access to items of interest on the Internet. Information
retrieval systems, such as Google, DevilFinder and Altavista have partially solved this
problem but prioritization and personalization (where a system maps available
content to user’s interests and preferences) of information were absent. This has
increased the demand for recommender systems more than ever before.
Having its emergence in mid of 1990s, recommendation engine is an information
filtering system uploading information tailored to users' interests, preferences, or
behavioural history on an item.
Practically, recommender systems encompass a class of techniques and algorithms
which are able to suggest “relevant” items to users. Items are ranked according to
their relevancy, and the most relevant ones are shown to the user. The relevancy is
something that the recommender system must determine and is mainly based on
historical data and users’ profile.
Recommender systems are beneficial to both service providers and users. With the
use of product recommendation systems, the customers are able to find the items
they are looking for easily and quickly. They are quickly becoming the primary way
for users to expose to the whole digital world through the lens of
their experiences, behaviours, preferences and interests. And in a world of
information density and product overload, a recommendation engine provides an
efficient way for companies to provide consumers with personalised
information and solutions.
This engine is a splendid marketing tool especially for e-commerce and is also useful
for increasing profits, sales and revenues in general. That's why personalized
product recommendations are so widely used in the retail industry, eleven more
highlighting the importance of recommendation engines in the e-commerce
industry.

 ARTIFICIAL INTELLIGENCE IN RECOMMENDATION

SYSTEM
In order to provide the user with service or product recommendations,
recommendation engines uses algorithm. Lately, they have started using machine
learning algorithms making the predicting process of items more accurate. Based on
the data received from recommendation systems, the algorithms change. The
collected data can be related to the explicit interactions i.e., information about the
users’ past activity, their ratings, reviews and other information about their profile,
such as gender, age or to the implicit interaction i.e., device the user uses for access,
clicks on a link, location, and dates.
Then the machine learning algorithm uses one of these techniques to sort the data
collected:
1. Content-based filtering
2. Collaborative filtering

1. Content-based filtering:
Content-based filtering is based on a single user’s interactions and preference. In
content-based filtering technique, recommendation is made based on the user
profiles using features extracted from the content of the items the user has
evaluated in the past. For example, recommendations will be based on looking at
established patterns in a user’s choice or behaviours. Returning information such
as products or services will relate to your likes or views.

Machine Learning Cycle- Content-based Filtering

Firstly the data is collected, which ranges from item feature and user profile data.
Item feature is based on specific content or label of the item for e.g. The
category, the main actors, the duration or other characteristics for the movies
(items). Whereas, user profile data ranges from the age, the sex, the job to any
other personal information of the user.
Once the raw data is processed and prepared there are many different
algorithms that can be applied to it.
In cluster analysis, it groups data object based on the information that describes
the object and their relationships. Another one is neural network that can be
trained to protect ratings or interaction based on the item and user attributes
 we can use deep neural nets to predict next action based on historical action and
content.
Finally the algorithm is optimized to get the desired result.
2. Collaborative filtering:
This approach makes recommendations based on other users with similar tastes
or situations. For example, by using their opinion and actions to recommend
items to you or to identify how one product may go well with another. ‘Next buy’
recommendations are a typical usage.
The essence of collaborative filtering is the following; two users who have liked
the same item before will like the one in the future.

AI

Machine Learning Cycle- Collaborative Filtering

The first step in the collaborative filtering using machine learning cycle is to
understand the data source. In this case the generic user behaviour serves from
all users together with the activities and preferences that they have and then in
the next step data selection is done in order to clean and transform it so that it
can be processed by the algorithm
There are two key algorithms for collaborative filtering namely user-user
collaborative filtering and item- item collaborative filtering.
In user- user filtering it searches for look-alike customer and offers product
based on what the look alike has chosen
In the second main algorithm, item- item collaborative filtering which rather
than finding a look-alike customer, finds look-alike products.
Once the multiple algorithms are applied which can be more than stated here,
they are compared and the impact and result are measured. This can be done by
measuring the increase in the revenue or by the increase the number of views.
This cycle can be repeated until the results are to an acceptable standard.
 3. Hybrid filtering:
Hybrid filtering technique combines different recommendation techniques in
order to gain better system optimization to avoid some limitations and problems
of pure recommendation systems. The idea behind hybrid techniques is that a
combination of algorithms will provide more accurate and effective
recommendations than a single algorithm as the disadvantages of one algorithm
can be overcome by another algorithm. For e.g. Netflix.

Though these are the most common types of recommender systems, there are
many other types and hybrids that are occasionally used.

 EXAMPLE OF RECOMMENDATION SYSTEM

During the last few decades, with the rise of YouTube, Amazon, Netflix and many
other such web services, recommender systems have taken more and more place in
our lives. From e-commerce (suggest to buyers articles that could interest them) to
online advertisement (suggest to users the right contents, matching their
preferences), recommender systems are today unavoidable in our daily online
journeys.
A vivid example of recommendation engine usage is by Amazon- with its “Customer
who bought this item also bought…”.Generally, the content recommendation
engine is like a clever and experienced salesperson who grasps the needs, tastes,
and requirements of the user and is capable to make knowledgeable decisions
about recommendations beneficial and relevant to the client's wants, meanwhile
increasing the conversion rate.
Working of Amazon:
Amazon keeps the corresponding items before the eyes of the customers by using
their browsing history. It provides the recommended and best-selling option based
on the usage of customer reviews and ratings. Actually, Amazon is inclined to sell
you a package instead of a product for e.g. a user has bought an earring, then it will
suggest them a corresponding necklace and bracelet.
It also uses recommendations for targeted marketing via email campaigns and
website pages. Hence, Amazon starts recommending a lot of products from
different categories based on the users’ browsing history and picks up those items
that they are probably going to buy.

Recommendation engines started their journey in e-commerce, nonetheless, they

are gaining more and more popularity in other spheres as well, such as in the Media.
A good example of recommendation engine usage in Media is done by YouTube and
Netflix.
YouTube with its “Recommended Videos” and “Other Movies You May Enjoy” by
Netflix are vivid examples of AI recommendation engine usage.
Netflix which uses recommendation system as its core. More than 80% of the TV
shows that people watch on the platform are discovered through a recommendation
system. That means the majority of what you decide to watch on Netflix is the result
of decisions made by a mysterious, black box of an algorithm. As a proof of the
importance of recommender systems, a few years ago, Netflix organised a challenges
(the “Netflix prize”) where the goal was to produce a recommender system that
performs better than its own algorithm with a prize of 1 million dollars to win.
Working of Netflix:
Netflix uses machine learning and algorithms whose aim is to help break viewers
preconceived notions and find shows that they might have chosen initially.
Its recommendation engine uses ‘three-legged stool’ working concept. The first leg is
the history of what Netflix members have watched (what people watch, what they
watch after, what they watch before, what they watched a year ago, what they’ve
watched recently and what time of day). Tags are done by Netflix employees who
understand everything about the content and proprietary machine learning
algorithms that take all of the data and put things together.
The latter – the second leg of the stool – is gathered from dozens of in-house and
freelance staff who watch every minute or every show on Netflix and tag it. The tags
they use range massively from how cerebral the piece is; to whether it has an
ensemble cast, is set in space, or stars a corrupt cop.
Then all of these tags and the user behaviour data are combined and a very
sophisticated machine learning algorithms is used to figure out what’s most
important and appropriate.
What those three things create for Netflix is ‘taste communities’ around the world.
It’s about people who watch the same kind of things that the user watches. Viewers
fit into multiple taste groups of which there are "a couple of thousand" – and it’s
these that affect what recommendations pop up to the top of your onscreen
interface, which genre rows are displayed, and how each row is ordered for each
individual viewer.
 BENEFITS OF AI IN RECOMMENDATION SYSTEM
 Improve customer satisfaction:
As the system is able to provide the user with more meaningful contents
whether these are product songs and videos
 Provide personalization:
The big benefit is that personalization level increase massively. This is
because you built a profile of the customer and therefore the system gets the
opportunity to learn their customers and provide them with personalized
user experience.
 Improve product discovery:
Because of this personalization the product discovery will improve
dramatically as the users might be finding products that they normally
wouldn't search for.
 Increased Sales:
An improved user experience leads to the increase in sales. According to the
studies McKinsey and techies merchants, they have found that the
recommendation system brought amazon 35% of its revenue and 23.7%
growth to best buy. In the streaming services, it brought up to 75% of video
consumption on Netflix and 60% of viewers on YouTube.

 COMMAN CHALLENGES
 Sparsity of data:
Data sets filled with rows and rows of values that contain blanks or zero
values. So finding ways to use denser parts of the data set and those with
information is critical.
 Latent association:
Labelling is imperfect. Same products with different labelling can be ignored
or incorrectly consumed, meaning that the information does not get
incorporated correctly.
 Scalability:
The traditional approach has become overwhelmed by the multiplicity of
products and clients. This becomes a challenge as data sets widen and can
lead to performance reduction.
 Synonymy:
It is the tendency of very similar items to have different names or entries.
Most recommender systems find it difficult to make distinction between
closely related items such as the difference between e.g. baby wear and baby
cloth.
 FUTURE WORK
Recommender systems research has incorporated a wide variety of Artificial
Intelligence (AI) techniques including machine learning, data mining, user
modelling, and case-based reasoning, among others.
Yet, there are many unaddressed issues and free rooms to be improved using
advanced techniques. For example, deep learning is a promising alternative to the
conventional neural networks. Deep learning methods are superior for effective
feature learning, especially when there is no known effective feature.
In today’s market where values are created by users’ networks, product and
service providers can no longer compete by simply comparing features and prices
to gain the competitive advantage. Recommender systems can be empowered
with theoretical foundations such as network effect, two-sided market, and game
theory. In fact, designing recommender systems as online two-sided platforms will
provide a rich source of data (ratings, items, users, etc.) and will open the door to
innovations in terms of dealing with data as a commodity. Moreover, the main
challenge is to determine the type and amount of incentives to provide to each
side (users and providers) so that the other side will follow.

 CONCLUSION
In conclusion, it is beyond doubt that the service engine has gained more
popularity and plays a significant role in the new digital era. In order to be
competitive in the market and get more efficient customers with
recommendation engines is in your best interests.
Especially with the use of artificial intelligence, in-the-moment recommendations
are more widespread, which is time-efficient and pragmatic. Thanks to artificial
intelligence, the recommendation engines have improved their productivity and
they are based on the customer's visual preferences rather than on the
description of the items.
 REFERENCES
 https://medium.com/voice-tech-podcast/a-simple-way-to-explain-the-
recommendation-engine-in-ai-d1a609f59d97
 https://www.kdnuggets.com/2019/09/machine-learning-recommender-
systems.html
 https://en.wikibooks.org/wiki/Artificial_Intelligence/Search/Recommender_S
ystems#Examples
 https://towardsdatascience.com/introduction-to-recommender-systems-
6c66cf15ada
 https://www.sciencedirect.com/science/article/pii/S1110866515000341#s00
50
 https://www.smarthint.co/en/ai-product-recommendation-
engine/#:~:text=Due%20to%20AI%2C%20recommendation%20engines,prefer
ences%20rather%20than%20product%20descriptions.
 https://www.wired.co.uk/article/how-do-netflixs-algorithms-work-machine-
learning-helps-to-predict-what-viewers-will-
like#:~:text=To%20help%20understand%2C%20consider%20a,vice%20preside
nt%20of%20product%20innovation.
 https://addepto.com/product-recommendation-system-algorithms-
challenges-benefits/
 https://medium.com/futuristone/the-future-directions-of-recommender-
systems-72beae7c2dd2
 https://www.youtube.com/watch?v=EjOlN6uVBOg
 https://www.youtube.com/watch?v=jhFFOmvPne4
 https://www.youtube.com/watch?v=F3ThOceGirE