Project Materials




Need help with a related project topic or New topic? Send Us Your Topic 




1.1 General Introduction
Intelligent systems require knowledge organisation to interpret, evaluate, and analyse acquired data. Most of our daily activities, such as e-commerce, online booking, social media, e-shopping, and other information-rich settings, necessitate the use of intelligent systems.

Recommender systems engage with users in a personalised manner, gathering information about a user’s tastes or preferences and using this knowledge to make suggestions and provide aid in circumstances when users must choose between a large number of viable solutions.

This chapter attempts to describe the recommender system and its techniques, as well as to introduce multi-criteria recommender systems and a genetic algorithm.

This chapter will also introduce the problem statement, goals and objectives, significance, and scope of this investigation.

1.2 Background Of The Study
When researchers began focusing on recommendation problems that obviously depend on rating structure in the mid-1990s, the recommender system was identified as a free study topic (Adomavicius & Tuzhilin, 2005).

Recommender systems (RSs) are methods and software tools for engaging with huge and complicated information spaces in order to prioritise and recommend products, offers, and objects that are likely to be of interest to a given user (Ricci, Rokach, & Shapira, 2015).

These recommendations relate to a variety of decision-making processes, such as which item or object to purchase, which movie to watch, which news to read online, which music to listen to, which airline to fly with, or which hotel to book (Ricci et al., 2015).

As a result of the variety of features of homogeneous products or services, related information, and choices available in the market place or in various application domains such as e-commerce, e-learning, e-government, and e-tourism, the recommender system has become widely used (Shambour, Hourani, & Fraihat, 2016).

The accuracy of the recommender system is an important component in determining how successfully it can acquire and process information. As a result, evaluating the recommender system has become an important and difficult undertaking.

One important aspect of evaluating the recommender 2 system is accuracy (Sohrabi, Toloo, Moeini, & Nalchigar, 2015). The goal of this research is to create an adaptive genetic algorithm to improve prediction accuracy and achieve a high correlation between expected and actual values in a multi-criteria recommender system.

1.2.1 Recommender Systems Techniques
The knowledge base, addressed domain, method, or technique employed during development can all affect the recommender system. Burke (2002) categorises the recommender into six distinct approaches:

The algorithm learns from the user’s previous likes and interests in this way, then offers matched things to the user based on that knowledge. The content-based approach relies on item features, thus a learning mechanism is used to establish the type of user profile that the content-based recommender will generate.

The similarity of the items is determined by the characteristics of the items (Ricci et al., 2015).
Collaborative filtering is the most well-known, established, and widely used technology (Burke, 2002).

It recommends things to the active user based on items previously enjoyed by other users with similar preferences. People-to-people correlation is used to describe collaborative filtering because the similarity in preferences of two users is reliant on their rating history (Ricci et al., 2015).

Demographic: The primary goal of this method is to categorise users based on personal characteristics and to recommend things based on a user’s demographic profile (Ricci et al., 2015). It may not be necessary to have a history of user ratings.

Information-based: This system suggests items based on a specific field of information about how beneficial an item is to the user and how various item features suit the user’s needs and preferences. The similarity metric might be viewed as the recommendation’s utility.

Community-based: This method suggests items to the user based on the preferences of the user’s friends. According to Ricci et al. (2015), people tend to rely more on the three suggestions of friends than on those of anonymous individuals with similar tastes.

Recommender systems that are hybrids: This type of recommender system combines two or more of the above-mentioned recommendation algorithms (Adomavicius & Tuzhilin, 2005).

A hybrid system that combines two methodologies attempts to harness the advantages of one to overcome the disadvantages of the other.

1.2.2  Multi-criteria Recommendation System
Traditionally, most RSs get the user’s overall or general preference for a certain item. In other words, it suggests things based on a single criterion rating by users, which is used as input information by the RS algorithm to evaluate user preference opinions.

Because users can express their thoughts based on some specific qualities of the item, a single criterion rating may give suggestions that do not fit the demands of the user in most circumstances.

Multi-criteria RSs, on the other hand, allow users to declare their choices for an item based on numerous attributes (Ricci et al., 2015). Multi-criteria ratings provide more information on the user’s preferences for numerous significant elements or components of an item (Adomavicius & Kwon, 2007).

The additional information on each user’s preferences will result in more accurate recommendations and higher recommendation quality.

Several recommender systems have incorporated multi-criteria ratings in recent years, rather than the usual single criterion evaluations (Ricci et al., 2015).

The goal of multi-criteria recommender systems is to take a step towards more efficiently and exquisitely analysing and comprehending users’ interests and choices and giving them with optimal answers.

1.2.3 Genetic Algorithm
Several computer scientists independently investigated evolving systems in the 1950s and 1960s, with the concept that evolution may be utilised as an optimisation technique for engineering challenges.

The goal was to use operators inspired by natural genetic variation and natural selection to evolve a population of candidate solutions to a given problem (Mitchell, 2004).
A genetic algorithm is a stochastic evolutionary search tool used for optimisation and learning. It is also a search method based on the idea for shifting from one population of “chromosomes” to a new population by combining natural selection (survival of the fittest) and natural genetics to solve an optimisation problem.

A genetic algorithm is also an evolutionary way to tackling optimisation problems including sequencing, travelling, salesman problems, and scheduling (Schmitt, 2001).

There are certain critical components to consider when developing a genetic algorithm, such as:Individuals are represented in a variety of ways, including bit strings, binary numbers, and real numbers.

Fitness function: Are you interested in the measurement of performance, which can be minimised or maximised?
Population: This contains a depiction of potential solutions.

Parent selection mechanism: Assists in distinguishing individuals based on their quality, allowing the better individual to become the next generation’s parent.

Variation operators: These operators generate new people from existing ones. They are classified as crossover (single or two point), which is performed on selected individuals to mix the general information to produce new individuals or children, and mutation (flipping).

The selection mechanism, also known as replacement, is predicated on survival of the fittest.
Because the genetic algorithm is stochastic and frequently guarantees no optimum solution, a suitable termination condition, such as when the fitness evaluation hits a certain limit, is required.

1.3 Statement Of The Problem
To evaluate users’ thoughts on experienced objects, the majority of current RSs use an overall estimation of a user rating of an item or single criterion rating methodologies.

Because the suitability of a recommended item for a specific user may be dependent on several important aspects or attributes in the user’s decision making, the traditional single criterion rating can be considered limited and inaccurate, because it cannot justify for the various items’ attributes.

A multi-criteria recommendation is proposed for this purpose, which integrates user ratings on many or diverse aspects of an item using an aggregate function-based technique.

To obtain a more accurate and economical prediction, the suggested technique employs an adaptive evolutionary algorithm to gain an appropriate learning connection.

1.4 Aims and Objectives Of The study
The goal of this research is to model a multi-criteria recommendation problem using an adaptive evolutionary algorithm and an aggregation function-based technique to create a more accurate and efficient forecast.

The precise goals were as follows: develop an adaptive genetic algorithm model; and apply an adaptive genetic algorithm to model multi-criteria recommendation problems.

To create a system that is capable of recommending the most suited item to a user.
To evaluate the predicted performance of the adaptive genetic algorithm-based multi-criteria recommender technique with the classic recommender approach.

1.5 ISignificance Of The Research
Web users and application domains such as e-commerce, e-learning, e-government, social networks, and e-tourism will benefit the most from this study because the system will make decision-making easier, faster, and more efficient.

A user rating of a multi-attribute item based on personal interest can significantly increase the prediction accuracy of the suggestion to other users.
1.6 Scope Of The Study
The study’s scope focuses on RSs, with a focus on multi-criteria RSs that will create a competent, accurate prediction for consumers.

This research comprises the creation of a complex system capable of recommending the best idea or item to consumers depending on their preferences.

1.7 Expected Results
The project’s goal is to evaluate the prediction performance of the proposed strategy to that of existing methodologies.

These results include lower prediction errors, higher ranking accuracy, and a good correlation between expected and actual values.

1.8 Thesis Structure
The remainder of this thesis is structured as follows: The second chapter provides an overview of the RS and multi-criteria RS, examines the adaptive genetic algorithm and its components, and evaluates related research.

The techniques and architecture of the study are described in Chapter 3. The system’s detailed implementation is presented in Chapter 4.

It also discusses the collected outcomes. Chapter 5 concludes with a discussion of the summary, conclusion, and recommendation.

Do you like to read the entire paper titled Enhancing Prediction Accuracy Of A Multi-Criteria Recommender System Using Adaptive Genetic Algorithm? Click below.

Need help with a related project topic or New topic? Send Us Your Topic 


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.