scholarly journals The Simpson’s Paradox in the Offline Evaluation of Recommendation Systems

2022 ◽  
Vol 40 (1) ◽  
pp. 1-22
Author(s):  
Amir H. Jadidinejad ◽  
Craig Macdonald ◽  
Iadh Ounis
Keyword(s):  
Recommender Systems ◽  
Recommendation System ◽  
Recommendation Systems ◽  
Rank Correlation ◽  
Open Loop ◽  
Evaluation Methodology ◽  
Simpson’S Paradox ◽  
Kendall Rank Correlation ◽  
Loop Feedback ◽  
Simpson's Paradox

Recommendation systems are often evaluated based on user’s interactions that were collected from an existing, already deployed recommendation system. In this situation, users only provide feedback on the exposed items and they may not leave feedback on other items since they have not been exposed to them by the deployed system. As a result, the collected feedback dataset that is used to evaluate a new model is influenced by the deployed system, as a form of closed loop feedback. In this article, we show that the typical offline evaluation of recommender systems suffers from the so-called Simpson’s paradox. Simpson’s paradox is the name given to a phenomenon observed when a significant trend appears in several different sub-populations of observational data but disappears or is even reversed when these sub-populations are combined together. Our in-depth experiments based on stratified sampling reveal that a very small minority of items that are frequently exposed by the deployed system plays a confounding factor in the offline evaluation of recommendation systems. In addition, we propose a novel evaluation methodology that takes into account the confounder, i.e., the deployed system’s characteristics. Using the relative comparison of many recommendation models as in the typical offline evaluation of recommender systems, and based on the Kendall rank correlation coefficient, we show that our proposed evaluation methodology exhibits statistically significant improvements of 14% and 40% on the examined open loop datasets (Yahoo! and Coat), respectively, in reflecting the true ranking of systems with an open loop (randomised) evaluation in comparison to the standard evaluation.

scholarly journals Research on Recommendation Systems using Deep Learning Models

2019 ◽  
Vol 8 (4) ◽  
pp. 10544-10551
Keyword(s):  
Deep Learning ◽  
Recommender Systems ◽  
Recommendation System ◽  
Recommendation Systems ◽  
Learning Models ◽  
Machine Learning Technique ◽  
Learning Technique ◽  
Future Preference ◽  
Movie Recommendation ◽  
News Recommendation

Recommender System is the effective tools that are accomplished of recommending the future preference of a set of products to the consumer and to predict the most likelihood items. Today, customers has the ability to purchase or sell different items with advancement of e-commerce website, nevertheless it made complicate to investigate the majority of appropriate items suitable for the interest of the consumer from many items. Due to this scenario, recommender systems that can recommend items appropriate for user's interest and likings have become mandatory. In recent days, various recommendation methods are applied to resolve the data abundance setback in numerous application areas like movie recommendation, e-commerce, news recommendation, song recommendation and social media. Even if all the available current recommender systems are successful in generating reasonable predictions, these recommendation system still facing the issues like accuracy, cold-start, sparsity and scalability problem. Deep learning, the recently developed sub domain of machine learning technique is utilized in recommendation systems to enhance the feature of predicted output. Deep Learning is used to generate recommendations and the research challenges specific to recommendation systems when using Deep Learning are also presented. In this research, the basic terminologies, the fundamental concepts of Recommendation engine and a wide-ranging review of deep learning models utilized in Recommender Systems are presented.

A review of a recommendation filtering system approach based on reliable sustainable opinion mining

2022 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Imen Gmach ◽  
Nadia Abaoub ◽  
Rubina Khan ◽  
Naoufel Mahfoudh ◽  
Amira Kaddour
Keyword(s):  
Collaborative Filtering ◽  
Recommender Systems ◽  
Information Search ◽  
Recommendation System ◽  
Opinion Mining ◽  
Information Filtering ◽  
Recommendation Systems ◽  
Content Filtering ◽  
Content Type ◽  
The Impact

PurposeIn this article the authors will focus on the state of the art on information filtering and recommender systems based on trust. Then the authors will represent a variety of filtering and recommendation techniques studied in different literature, like basic content filtering, collaborative filtering and hybrid filtering. The authors will also examine different trust-based recommendation algorithms. It will ends with a summary of the different existing approaches and it develops the link between trust, sustainability and recommender systems.Design/methodology/approachMethodology of this study will begin with a general introduction to the different approaches of recommendation systems; then define trust and its relationship with recommender systems. At the end the authors will present their approach to “trust-based recommendation systems”.FindingsThe purpose of this study is to understand how groups of users could improve trust in a recommendation system. The authors will examine how to evaluate the performance of recommender systems to ensure their ability to meet the needs that led to its creation and to make the system sustainable with respect to the information. The authors know very well that selecting a measure must depend on the type of data to be processed and user interests. Since the recommendation domain is derived from information search paradigms, it is obvious to use the evaluation measures of information systems.Originality/valueThe authors presented a list of recommendations systems. They examined and compared several recommendation approaches. The authors then analyzed the dominance of collaborative filtering in the field and the emergence of Recommender Systems in social web. Then the authors presented and analyzed different trust algorithms. Finally, their proposal was to measure the impact of trust in recommendation systems.

scholarly journals Recommendation Systems: Algorithms, Challenges, Metrics, and Business Opportunities

2020 ◽  
Vol 10 (21) ◽  
pp. 7748
Author(s):  
Zeshan Fayyaz ◽  
Mahsa Ebrahimian ◽  
Dina Nawara ◽  
Ahmed Ibrahim ◽  
Rasha Kashef
Keyword(s):  
Recommender Systems ◽  
Recommendation System ◽  
Information Overload ◽  
State Of The Art ◽  
Recommendation Systems ◽  
Qualitative Evaluation ◽  
Potential Influence ◽  
Systems Research ◽  
Current State

Recommender systems are widely used to provide users with recommendations based on their preferences. With the ever-growing volume of information online, recommender systems have been a useful tool to overcome information overload. The utilization of recommender systems cannot be overstated, given its potential influence to ameliorate many over-choice challenges. There are many types of recommendation systems with different methodologies and concepts. Various applications have adopted recommendation systems, including e-commerce, healthcare, transportation, agriculture, and media. This paper provides the current landscape of recommender systems research and identifies directions in the field in various applications. This article provides an overview of the current state of the art in recommendation systems, their types, challenges, limitations, and business adoptions. To assess the quality of a recommendation system, qualitative evaluation metrics are discussed in the paper.

scholarly journals Movie Recommended System by Using Collaborative Filtering

2019 ◽  
pp. 9-15
Author(s):  
Bheema Shireesha ◽  
Navuluri Madhavilatha ◽  
Chunduru Anilkumar
Keyword(s):  
Decision Making ◽  
Collaborative Filtering ◽  
Recommender Systems ◽  
Recommendation System ◽  
Recommendation Systems ◽  
Large Size

Recommendation system helps people in decision making an item/person. Recommender systems are now pervasive and seek to make profit out of customers or successfully meet their needs. Companies like Amazon use their huge amounts of data to give recommendations for users. Based on similarities among items, systems can give predictions for a new item’s rating. Recommender systems use the user, item, and ratings information to predict how other users will like a particular item. In this project, we attempt to under- stand the different kinds of recommendation systems and compare their performance on the Movie Lens dataset. Due to large size of data, recommendation system suffers from scalability problem. Hadoop is one of the solutions for this problem.

scholarly journals Implementing Recommender Systems using Machine Learning and Knowledge Discovery Tools

2021 ◽  
Vol 2 (2) ◽  
pp. 44-53
Author(s):  
Mohammad Zahrawi ◽  
Ahmad Mohammad
Keyword(s):  
Machine Learning ◽  
Knowledge Discovery ◽  
Recommender Systems ◽  
Learning Strategies ◽  
Search Engines ◽  
Recommendation System ◽  
Recommendation Systems ◽  
Relevant Information ◽  
Different Characteristics ◽  
Better Than

The current research offers a novel use of machine learning strategies to create a recommendation system. At recent era, recommender systems (RSs) have been used widely in e-commerce, entertainment purposes, and search engines. In more general, RSs are set of algorithms designed to recommend relevant items to users (movies to watch, books to read, products to buy, songs to listen, and others). This article discovers the different characteristics and features of many approaches used for recommendation systems in order to filter and prioritize the relevant information and work as a compass for searching. Recommender engines are crucial in some companies as they can create a big amount of income when they are effective or be a way to stand out remarkably from other rivals. As a proof of the importance of recommender engine, it can be stated that Netflix arrange a challenge (the “Netflix prize”) where the mission was to create a recommender engine that achieves better than its own algorithm with a prize of 1 million dollars to win.

scholarly journals Are Scientific Models of Life Testable? A Lesson from Simpson’s Paradox

Sci ◽  
2020 ◽  
Vol 2 (3) ◽  
pp. 73
Author(s):  
Prasanta S. Bandyopadhyay ◽  
Nolan Grunska ◽  
Don Dcruz ◽  
Mark C. Greenwood
Keyword(s):  
Origin Of Life ◽  
Rna World ◽  
Scientific Models ◽  
Scientific Model ◽  
Simpson’S Paradox ◽  
World Theory ◽  
Life Issues ◽  
The Origin Of Life ◽  
Logical Sense ◽  
Simpson's Paradox

We address the need for a model by considering two competing theories regarding the origin of life: (i) the Metabolism First theory and (ii) the RNA World theory. We discuss two inter-related points. (I) Models are valuable tools in understanding both the processes and intricacies of the origin of life issues. (II) Insights from models also help us to evaluate the core objection to origin of life theories called “the inefficiency objection” commonly raised by proponents of both the Metabolism First theory and the RNA World theory against each other. We use Simpson’s paradox as a tool for challenging this objection. We will use models in various senses ranging from taking them as representations of reality to treating them as theories/accounts that provide heuristics for probing reality. In this paper, we will frequently use models and theories interchangeably. Additionally, we investigate Conway’s Game of Life and contrast it with our Simpson’s Paradox (SP)-based approach to emergence of life issues. Finally, we discuss some of the consequences of our view. A scientific model is testable in three senses: (i) a logical sense, (ii) a nomological sense, and (iii) a current technological sense. The SP-based model is testable in the logical sense. It is also testable nomologically. However, it is not currently feasible to test it.

scholarly journals Fashion Recommendation Systems, Models and Methods: A Review

Informatics ◽  
2021 ◽  
Vol 8 (3) ◽  
pp. 49
Author(s):  
Samit Chakraborty ◽  
Md. Saiful Hoque ◽  
Naimur Rahman Jeem ◽  
Manik Chandra Biswas ◽  
Deepayan Bardhan ◽  
...  
Keyword(s):  
Recommendation System ◽  
Recommendation Systems ◽  
Fast Fashion ◽  
Potential Models ◽  
Scholarly Article ◽  
Enormous Amount ◽  
Shopping Experience ◽  
Tremendous Amount ◽  
Systems Models ◽  
Filtering Techniques

In recent years, the textile and fashion industries have witnessed an enormous amount of growth in fast fashion. On e-commerce platforms, where numerous choices are available, an efficient recommendation system is required to sort, order, and efficiently convey relevant product content or information to users. Image-based fashion recommendation systems (FRSs) have attracted a huge amount of attention from fast fashion retailers as they provide a personalized shopping experience to consumers. With the technological advancements, this branch of artificial intelligence exhibits a tremendous amount of potential in image processing, parsing, classification, and segmentation. Despite its huge potential, the number of academic articles on this topic is limited. The available studies do not provide a rigorous review of fashion recommendation systems and the corresponding filtering techniques. To the best of the authors’ knowledge, this is the first scholarly article to review the state-of-the-art fashion recommendation systems and the corresponding filtering techniques. In addition, this review also explores various potential models that could be implemented to develop fashion recommendation systems in the future. This paper will help researchers, academics, and practitioners who are interested in machine learning, computer vision, and fashion retailing to understand the characteristics of the different fashion recommendation systems.

Legal document recommendation system: A cluster based pairwise similarity computation

2021 ◽  
pp. 1-13
Author(s):  
Jenish Dhanani ◽  
Rupa Mehta ◽  
Dipti Rana
Keyword(s):  
Recommender Systems ◽  
Recommendation System ◽  
Real Life ◽  
Citation Network ◽  
Search Space ◽  
Pairwise Similarity ◽  
Large Numbers ◽  
Legal Document ◽  
Legal Domain ◽  
Similarity Scores

Legal practitioners analyze relevant previous judgments to prepare favorable and advantageous arguments for an ongoing case. In Legal domain, recommender systems (RS) effectively identify and recommend referentially and/or semantically relevant judgments. Due to the availability of enormous amounts of judgments, RS needs to compute pairwise similarity scores for all unique judgment pairs in advance, aiming to minimize the recommendation response time. This practice introduces the scalability issue as the number of pairs to be computed increases quadratically with the number of judgments i.e., O (n2). However, there is a limited number of pairs consisting of strong relevance among the judgments. Therefore, it is insignificant to compute similarities for pairs consisting of trivial relevance between judgments. To address the scalability issue, this research proposes a graph clustering based novel Legal Document Recommendation System (LDRS) that forms clusters of referentially similar judgments and within those clusters find semantically relevant judgments. Hence, pairwise similarity scores are computed for each cluster to restrict search space within-cluster only instead of the entire corpus. Thus, the proposed LDRS severely reduces the number of similarity computations that enable large numbers of judgments to be handled. It exploits a highly scalable Louvain approach to cluster judgment citation network, and Doc2Vec to capture the semantic relevance among judgments within a cluster. The efficacy and efficiency of the proposed LDRS are evaluated and analyzed using the large real-life judgments of the Supreme Court of India. The experimental results demonstrate the encouraging performance of proposed LDRS in terms of Accuracy, F1-Scores, MCC Scores, and computational complexity, which validates the applicability for scalable recommender systems.

Sign in / Sign up

Export Citation Format

Share Document