Content-based top-N recommendations with perceived similarity

Wang, Charlie C. L.; Agrawal, Arpita; Li, Xiaojun; Makkad, Tanima; Veljee, Ejaz; Mengshoel, Ole J.; Jude, Alvin

doi:10.1109/smc.2017.8122750

Cited by 7 publications

(5 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Their experiments showed that users indeed preferred the recommendations that considered the human similarity judgments. The results by Colucci et al (2016) and Wang et al (2017) indicate that humans largely agree in their similarity perceptions of item pairs, while these perceptions are not aligned with algorithmic similarity functions. Further, their works demonstrate that different similarity functions result in discrepancies between objective similarity measures and human perception.…”

Section: Assessing the Similarity Perception Of Item Pairsmentioning

confidence: 87%

“…Yet, these human judgments were only partly aligned with the output of three algorithmic similarity functions (with the highest precision value being only .55). Building on this dataset, Wang et al (2017) designed two content-based recommendation approaches, where one considered human perceptions in the recommendation process whereas the other did not. Their experiments showed that users indeed preferred the recommendations that considered the human similarity judgments.…”

Section: Assessing the Similarity Perception Of Item Pairsmentioning

confidence: 99%

See 1 more Smart Citation

Intra-list similarity and human diversity perceptions of recommendations: the details matter

Bauer

Jannach

2022

User Model User-Adap Inter

View full text Add to dashboard Cite

The diversity of the generated item suggestions can be an important quality factor of a recommender system. In offline experiments, diversity is commonly assessed with the help of the intra-list similarity (ILS) measure, which is defined as the average pairwise similarity of the items in a list. The similarity of each pair of items is often determined based on domain-specific meta-data, e.g., movie genres. While this approach is common in the literature, it in most cases remains open if a particular implementation of the ILS measure is actually a valid proxy for the human diversity perception in a given application. With this work, we address this research gap and investigate the correlation of different ILS implementations with human perceptions in the domains of movie and recipe recommendation. We conducted several user studies involving over 500 participants. Our results indicate that the particularities of the ILS metric implementation matter. While we found that the ILS metric can be a good proxy for human perceptions, it turns out that it is important to individually validate the used ILS metric implementation for a given application. On a more general level, our work points to a certain level of oversimplification in recommender systems research when it comes to the design of computational proxies for human quality perceptions and thus calls for more research regarding the validation of the corresponding metrics.

show abstract

Section: Assessing the Similarity Perception Of Item Pairsmentioning

confidence: 87%

Section: Assessing the Similarity Perception Of Item Pairsmentioning

confidence: 99%

Intra-list similarity and human diversity perceptions of recommendations: the details matter

Bauer

Jannach

2022

User Model User-Adap Inter

View full text Add to dashboard Cite

show abstract

“…A problem with the validation is that since every y cnd is a new checklist, the ground truths l needed to evaluate y cnd can be missing. A common solution to this problem is to collect the ground truth empirically [23], but this is not an option for us. To get a meaningful validation result, the performance statistics for the evaluation need to be estimated.…”

Section: Experiments 3: Evaluation Of Constructed Checklistsmentioning

confidence: 99%

Bayesian Feature Construction for Case-Based Reasoning: Generating Good Checklists

Flogard

Mengshoel

Bach

2021

Case-Based Reasoning Research and Development

View full text Add to dashboard Cite

Checklists are used to aid the fulfillment of safety critical activities in a variety of different applications, such as aviation, health care or labour inspections. However, optimizing a checklist for a specific purpose can be challenging. Checklists also need to be trustworthy and user friendly to promote user compliance. With labour inspections as a starting point, we introduce the Checklist Construction Problem. To address the problem, we seek to optimize the content of labour inspection checklists in order to improve the working conditions in every organisation targeted for inspections. To do so, we introduce a hybrid framework called BCBR to construct trustworthy checklists. BCBR is based on case-based reasoning (CBR) and Bayesian inference (BI) and constructs new checklists based on past cases. A key novelty of BCBR is the use of BI for constructing new features in past cases. The augmented past cases are retrieved via CBR to construct new checklists, which ensures justification for the content of the checklists and promotes trust. Experiments suggest that BCBR is more effective than any other baseline we tested, in terms of constructing trustworthy checklists.

show abstract

“…CB similarity could alternately be done with tags or keywords, contributed by users or domain experts; MovieLens released a set of user-contributed tags for movies via the Tag Genome Project [VSR12]. CB-similarity can also perform YML recommendations [CZG16; NK11], and improving CB similarity using supervised learning can improve YML recommendations [WAL17]. Both approaches have its own downsides; CF requires user data making it unsuitable for new items, while CB could produce only obvious recommendations.…”

Section: Related Workmentioning

confidence: 99%

Finding Similar Movies: Dataset, Tools, and Methods

Leng¹,

Paulino²,

Haider³

et al. 2018

CSRN

Self Cite

View full text Add to dashboard Cite

Recommender systems are becoming ubiquitous in online commerce as well as in video-on-demand (VOD) and music streaming services. A popular form of giving recommendations is to base them on a currently selected product (or items), and provide "More Like This," "Items Similar to This," or "People Who Bought This also Bought" functionality. These recommendations are based on similarity computations, also known as item-item similarity computations. Such computations are typically implemented by heuristic algorithms, which may not match the perceived item-item similarity of users. In contrast, we study in this paper a data-driven approach to similarity for movies using labels crowdsourced from a previous work. Specifically, we develop four similarity methods and investigate how user-contributed labels can be used to improve similarity computations to better match user perceptions in movie recommendations. These four methods were tested against the best known method with a user experiment (n = 114) using the MovieLens 20M dataset. Our experiment showed that all our supervised methods beat the unsupervised benchmark and the differences were both statistically and practically significant. This paper's main contributions include user evaluation of similarity methods for movies, user-contributed labels indicating movie similarities, and code for the annotation tool which can be found at http://MovieSim.org.

show abstract

Content-based top-N recommendations with perceived similarity

Cited by 7 publications

References 19 publications

Intra-list similarity and human diversity perceptions of recommendations: the details matter

Intra-list similarity and human diversity perceptions of recommendations: the details matter

Bayesian Feature Construction for Case-Based Reasoning: Generating Good Checklists

Finding Similar Movies: Dataset, Tools, and Methods

Contact Info

Product

Resources

About