Recommender System with Machine Learning and Artificial Intelligence

Practical Tools and Applications in Medical, Agricultural and Other Industries

Sachi Nandan Mohanty, Jyotir Moy Chatterjee, Sarika Jain, Ahmed A. Elngar, Priya Gupta, Sachi Nandan Mohanty, Jyotir Moy Chatterjee, Sarika Jain, Ahmed A. Elngar, Priya Gupta

• Publisher: Wiley-Scrivener
• Year: 2020
• ISBN:
  9781119711605
• Edition: 1
• Topic: Computer Science
• Subtopic: Data Warehousing
• Index: Computer Science

This book is a multi-disciplinary effort that involves world-wide experts from diverse fields, such as artificial intelligence, human computer interaction, information technology, data mining, statistics, adaptive user interfaces, decision support systems, marketing, and consumer behavior. It comprehensively covers the topic of recommender systems, which provide personalized recommendations of items or services to the new users based on their past behavior. Recommender system methods have been adapted to diverse applications including social networking, movie recommendation, query log mining, news recommendations, and computational advertising.

This book synthesizes both fundamental and advanced topics of a research area that has now reached maturity. Recommendations in agricultural or healthcare domains and contexts, the context of a recommendation can be viewed as important side information that affects the recommendation goals. Different types of context such as temporal data, spatial data, social data, tagging data, and trustworthiness are explored. This book illustrates how this technology can support the user in decision-making, planning and purchasing processes in agricultural & healthcare sectors.

Part I
INTRODUCTION TO RECOMMENDER SYSTEMS

1
An Introduction to Basic Concepts on Recommender Systems

Pooja Rana, Nishi Jain and Usha Mittal*

Department of Computer Science and Engineering, Lovely Professional University, Phagwara, India

Abstract

In today’s world, we find a wide range of possibilities of any search that we do online and we might find difficulties in choosing what we actually need. To address these issues, recommendation System plays a major role. A recommender system is a filtering system that filters the data using different algorithms and recommends the most relevant data to the user. For instance, a recommender system for e-commerce requires a past history of the site and if the user is not having any past history then the recommender system recommends the bestselling product or most popular product present in the market. Recommendation systems are effective tools for personalization, are always up-to-date, and gives a recommendation based on actual user behavior. Besides being useful in buying products it has a few disadvantages like it is difficult to set up and get running as they are database-driven. Sometimes recommendations are wrong which makes customers unsatisfied. Recommender system is used in different areas like recommendation for entertainment such as movies, songs etc., e-learning web site recommendation, newspaper recommendation and e-mail filters.

In this chapter, various recommendation techniques with their pros and cons and different evaluation metrices has been discussed.

Keywords: Recommendation, item-based, rating, artificial intelligence

1.1 Introduction

A recommender system is a sub-category of an information extraction system that helps to find the ranking or user preference for a particular item. Recommendation systems are dedicated software and methods that give ideas related to things that are used by different users [1, 2]. Many decisions can be made by considering recommendations like which product to purchase, type of music to listen, or what and where to read online news.

The things that are suggested by the system are known as “Item”. A recommender system generally concentrates on a particular form of item like DVDs, or articles and thus its proposal, its graphical user interface (GUI), and the primary method used to make the suggestions are adapted to give beneficial and real recommendations for a particular form of item.

Consider an example of toy recommendation system that assists customers to select a toy to buy. The popular e-commerce web site i.e. Amazon.com also uses a recommendation system to identify the online store for every user [3]. As suggestions and choices are generally personalized, different users or user groups get different suggestions. In case of magazines and newspapers, non-personalized suggestions are produced, which are very easy to generate. Consider the example of selecting best ten books or CDs.

To make the personalized suggestions, ranked lists of items are produced. User’s preferences and constrains are considered for generating the ranking to extract the most suitable products and services. For computing the most similar products and services, user’s preferences are collected implicitly or explicitly by understanding the users’ actions.

The fast development and diverse data existing on the web and existence of new e-business services like purchasing goods, comparing items, auction, etc. often stunned customers, leading them to make wrong choices. Thus, rather than giving benefit to customers, it starts decreasing the well-being. Actually, choice, with its insinuations of liberty, independence, and self-determination can become dangerous; generating a sense that independence may come to be observed as a kind of misery-inducing dictatorship [4].

These days, use of recommender systems has widely increased, indicated by the following facts:

1. In high-rated internet sites, recommendation systems have a vital role like Yahoo, Amazon.com, Netflix, YouTube, TripAdvisor and IMDB. Now, RSs as a part of service have been provided to the subscriber by many media companies. For instance, Netflix, the online movie rental service, has paid a great price i.e. one million dollars as prize to the group that first successfully improved the accuracy of its recommendation system [3].
2. ACM Recommender Systems (RecSys) has been established in 2007 dedicated for conferences and workshops.

1.2 Functions of Recommendation Systems

Recommender system offers suggestions to the user about a particular item that user wants to use. Now this definition can be refined by representing the different roles that a system can play. Recommender system plays different roles according to the user for example; a recommender system used by travel intermediary is usually used to increase the revenue like Expedia.com and Visitfinland.com while the customer’s objective for using the systems is to find an appropriate hotel and interesting events/attractions when visiting a destination.

The following are different reasons to exploit RS technology by service providers:

1. Increase the sale of product: The major objective of a commercial RS is to increase its sale, or in other words to sell those products also which can’t be sold without recommendations. Recommendations are provided considering that suggested products and services meet the customer’s requirements. Non-commercial recommendations are used for similar objectives. Consider an example of a content writer who wants to increase the number of news reader on his site. The goal of the service provider to use the recommender system is to increase the users that opts the products or services as compared to users surf the site.
2. Selling variety of products: RS also help a user to find items that might be difficult to find without a particular reference. For example, the recommender system used in Netflix has the goal of renting maximum movies in the list, rather than the most popular movies. Making such recommendations could be hard without a recommender system because the service provider cannot take the risk of suggesting videos which do not meet the user’s taste. In this way, the recommender system also suggests movies which are not even popular.
3. User satisfaction. The recommender system helps in improving the experience of the person with the application or web site. It provides interesting, significant and, relevant recommendations as well as provides better human– computer interaction. The effective recommendations i.e. accurate as well as interact...