Selecting appropriate tourist attractions to visit in real time is an important problem for travellers.Since recommenders proactively suggest items based on user preference,they are a promising solution for this probl...Selecting appropriate tourist attractions to visit in real time is an important problem for travellers.Since recommenders proactively suggest items based on user preference,they are a promising solution for this problem.Travellers visit tourist attractions sequentially by considering multiple attributes at the same time.Therefore,it is desirable to consider this when developing recommenders for tourist attractions.Using GRU4REC,we proposed RNN-based sequence-aware recommenders(RNN-SARs)that use multiple sequence datasets for training the recommended model,named multi-RNN-SARs.We proposed two types of multi-RNN-SARs-concatenate-RNN-SARs and parallel-RNN-SARs.In order to evaluate multi-RNN-SARs,we compared hit rate(HR)and mean reciprocal rank(MRR)of the item-based collaborative filtering recommender(item-CFR),RNN-SAR with the single-sequence dataset(basic-RNN-SAR),multi-RNN-SARs and the state-of-the-art SARs using a real-world travel dataset.Our research shows that multi-RNN-SARs have significantly higher performances compared to item-CFR.Not all multi-RNNSARs outperform basic-RNN-SAR but the best multi-RNN-SAR achieves comparable performance to that of the state-of-the-art algorithms.These results highlight the importance of using multiple sequence datasets in RNN-SARs and the importance of choosing appropriate sequence datasets and learning methods for implementing multi-RNN-SARs in practice.展开更多
A recommender system is a tool designed to suggest relevant items to users based on their preferences and behaviors.Collaborative filtering,a popular technique within recommender systems,predicts user interests by ana...A recommender system is a tool designed to suggest relevant items to users based on their preferences and behaviors.Collaborative filtering,a popular technique within recommender systems,predicts user interests by analyzing patterns in interactions and similarities between users,leveraging past behavior data to make personalized recommendations.Despite its popularity,collaborative filtering faces notable challenges,and one of them is the issue of grey-sheep users who have unusual tastes in the system.Surprisingly,existing research has not extensively explored outlier detection techniques to address the grey-sheep problem.To fill this research gap,this study conducts a comprehensive comparison of 12 outlier detectionmethods(such as LOF,ABOD,HBOS,etc.)and introduces innovative user representations aimed at improving the identification of outliers within recommender systems.More specifically,we proposed and examined three types of user representations:1)the distribution statistics of user-user similarities,where similarities were calculated based on users’rating vectors;2)the distribution statistics of user-user similarities,but with similarities derived from users represented by latent factors;and 3)latent-factor vector representations.Our experiments on the Movie Lens and Yahoo!Movie datasets demonstrate that user representations based on latent-factor vectors consistently facilitate the identification of more grey-sheep users when applying outlier detection methods.展开更多
This study introduces an advanced recommender system for technology enhanced learning(TEL)that synergizes neural collaborative filtering,sentiment analysis,and an adaptive learning rate to address the limitations of t...This study introduces an advanced recommender system for technology enhanced learning(TEL)that synergizes neural collaborative filtering,sentiment analysis,and an adaptive learning rate to address the limitations of traditional TEL systems.Recognizing the critical gap in existing approaches—primarily their neglect of user emotional feedback and static learning paths—our model innovatively incorporates sentiment analysis to capture and respond to nuanced emotional feedback from users.Utilizing bidirectional encoder representations from Transformers for sentiment analysis,our system not only understands but also respects user privacy by processing feedback without revealing sensitive information.The adaptive learning rate,inspired by AdaGrad,allows our model to adjust its learning trajectory based on the sentiment scores associated with user feedback,ensuring a dynamic response to both positive and negative sentiments.This dual approach enhances the system’s adapt-ability to changing user preferences and improves its contentment understanding.Our methodology involves a comprehensive analysis of both the content of learning materials and the behaviors and preferences of learners,facilitating a more personalized learning experience.By dynamically adjusting recommendations based on real-time user data and behavioral analysis,our system leverages the collective insights of similar users and rele-vant content.We validated our approach against three datasets-MovieLens,Amazon,and a proprietary TEL dataset—and saw significant improvements in recommendation precision,F-score,and mean absolute error.The results indicate the potential of integrating sentiment analysis and adaptive learning rates into TEL recommender systems,marking a step forward in developing more responsive and user-centric educational technologies.This study paves the way for future advancements in TEL systems,emphasizing the importance of emotional intelli-gence and adaptability in enhancing the learning experience.展开更多
Music recommendation systems are essential due to the vast amount of music available on streaming platforms,which can overwhelm users trying to find new tracks that match their preferences.These systems analyze users...Music recommendation systems are essential due to the vast amount of music available on streaming platforms,which can overwhelm users trying to find new tracks that match their preferences.These systems analyze users’emotional responses,listening habits,and personal preferences to provide personalized suggestions.A significant challenge they face is the“cold start”problem,where new users have no past interactions to guide recommendations.To improve user experience,these systems aimto effectively recommendmusic even to such users by considering their listening behavior and music popularity.This paper introduces a novel music recommendation system that combines order clustering and a convolutional neural network,utilizing user comments and rankings as input.Initially,the system organizes users into clusters based on semantic similarity,followed by the utilization of their rating similarities as input for the convolutional neural network.This network then predicts ratings for unreviewed music by users.Additionally,the system analyses user music listening behaviour and music popularity.Music popularity can help to address cold start users as well.Finally,the proposed method recommends unreviewed music based on predicted high rankings and popularity,taking into account each user’s music listening habits.The proposed method combines predicted high rankings and popularity by first selecting popular unreviewedmusic that themodel predicts to have the highest ratings for each user.Among these,the most popular tracks are prioritized,defined by metrics such as frequency of listening across users.The number of recommended tracks is aligned with each user’s typical listening rate.The experimental findings demonstrate that the new method outperformed other classification techniques and prior recommendation systems,yielding a mean absolute error(MAE)rate and rootmean square error(RMSE)rate of approximately 0.0017,a hit rate of 82.45%,an average normalized discounted cumulative gain(nDCG)of 82.3%,and a prediction accuracy of new ratings at 99.388%.展开更多
The performance of deep recommendation models degrades significantly under data poisoning attacks.While adversarial training methods such as Vulnerability-Aware Training(VAT)enhance robustness by injecting perturbatio...The performance of deep recommendation models degrades significantly under data poisoning attacks.While adversarial training methods such as Vulnerability-Aware Training(VAT)enhance robustness by injecting perturbations into embeddings,they remain limited by coarse-grained noise and a static defense strategy,leaving models susceptible to adaptive attacks.This study proposes a novel framework,Self-Purification Data Sanitization(SPD),which integrates vulnerability-aware adversarial training with dynamic label correction.Specifically,SPD first identifies high-risk users through a fragility scoring mechanism,then applies self-purification by replacing suspicious interactions with model-predicted high-confidence labels during training.This closed-loop process continuously sanitizes the training data and breaks the protection ceiling of conventional adversarial training.Experiments demonstrate that SPD significantly improves the robustness of both Matrix Factorization(MF)and LightGCN models against various poisoning attacks.We show that SPD effectively suppresses malicious gradient propagation and maintains recommendation accuracy.Evaluations on Gowalla and Yelp2018 confirmthat SPD-trainedmodels withstandmultiple attack strategies—including Random,Bandwagon,DP,and Rev attacks—while preserving performance.展开更多
Recommendation systems have become indispensable for providing tailored suggestions and capturing evolving user preferences based on interaction histories.The collaborative filtering(CF)model,which depends exclusively...Recommendation systems have become indispensable for providing tailored suggestions and capturing evolving user preferences based on interaction histories.The collaborative filtering(CF)model,which depends exclusively on user-item interactions,commonly encounters challenges,including the cold-start problem and an inability to effectively capture the sequential and temporal characteristics of user behavior.This paper introduces a personalized recommendation system that combines deep learning techniques with Bayesian Personalized Ranking(BPR)optimization to address these limitations.With the strong support of Long Short-Term Memory(LSTM)networks,we apply it to identify sequential dependencies of user behavior and then incorporate an attention mechanism to improve the prioritization of relevant items,thereby enhancing recommendations based on the hybrid feedback of the user and its interaction patterns.The proposed system is empirically evaluated using publicly available datasets from movie and music,and we evaluate the performance against standard recommendation models,including Popularity,BPR,ItemKNN,FPMC,LightGCN,GRU4Rec,NARM,SASRec,and BERT4Rec.The results demonstrate that our proposed framework consistently achieves high outcomes in terms of HitRate,NDCG,MRR,and Precision at K=100,with scores of(0.6763,0.1892,0.0796,0.0068)on MovieLens-100K,(0.6826,0.1920,0.0813,0.0068)on MovieLens-1M,and(0.7937,0.3701,0.2756,0.0078)on Last.fm.The results show an average improvement of around 15%across all metrics compared to existing sequence models,proving that our framework ranks and recommends items more accurately.展开更多
Recommendation systems are key to boosting user engagement,satisfaction,and retention,particularly on media platforms where personalized content is vital.Sequential recommendation systems learn from user-item interact...Recommendation systems are key to boosting user engagement,satisfaction,and retention,particularly on media platforms where personalized content is vital.Sequential recommendation systems learn from user-item interactions to predict future items of interest.However,many current methods rely on unique user and item IDs,limiting their ability to represent users and items effectively,especially in zero-shot learning scenarios where training data is scarce.With the rapid development of Large Language Models(LLMs),researchers are exploring their potential to enhance recommendation systems.However,there is a semantic gap between the linguistic semantics of LLMs and the collaborative semantics of recommendation systems,where items are typically indexed by IDs.Moreover,most research focuses on item representations,neglecting personalized user modeling.To address these issues,we propose a sequential recommendation framework using LLMs,called CIT-Rec,a model that integrates Collaborative semantics for user representation and Image and Text information for item representation to enhance Recommendations.Specifically,by aligning intuitive image information with text containing semantic features,we can more accurately represent items,improving item representation quality.We focus not only on item representations but also on user representations.To more precisely capture users’personalized preferences,we use traditional sequential recommendation models to train on users’historical interaction data,effectively capturing behavioral patterns.Finally,by combining LLMs and traditional sequential recommendation models,we allow the LLM to understand linguistic semantics while capturing collaborative semantics.Extensive evaluations on real-world datasets show that our model outperforms baseline methods,effectively combining user interaction history with item visual and textual modalities to provide personalized recommendations.展开更多
Recommending personalized travel routes from sparse,implicit feedback poses a significant challenge,as conventional systems often struggle with information overload and fail to capture the complex,sequential nature of...Recommending personalized travel routes from sparse,implicit feedback poses a significant challenge,as conventional systems often struggle with information overload and fail to capture the complex,sequential nature of user preferences.To address this,we propose a Conditional Generative Adversarial Network(CGAN)that generates diverse and highly relevant itineraries.Our approach begins by constructing a conditional vector that encapsulates a user’s profile.This vector uniquely fuses embeddings from a Heterogeneous Information Network(HIN)to model complex user-place-route relationships,a Recurrent Neural Network(RNN)to capture sequential path dynamics,and Neural Collaborative Filtering(NCF)to incorporate collaborative signals from the wider user base.This comprehensive condition,further enhanced with features representing user interaction confidence and uncertainty,steers a CGAN stabilized by spectral normalization to generate high-fidelity latent route representations,effectively mitigating the data sparsity problem.Recommendations are then formulated using an Anchor-and-Expand algorithm,which selects relevant starting Points of Interest(POI)based on user history,then expands routes through latent similarity matching and geographic coherence optimization,culminating in Traveling Salesman Problem(TSP)-based route optimization for practical travel distances.Experiments on a real-world check-in dataset validate our model’s unique generative capability,achieving F1 scores ranging from 0.163 to 0.305,and near-zero pairs−F1 scores between 0.002 and 0.022.These results confirm the model’s success in generating novel travel routes by recommending new locations and sequences rather than replicating users’past itineraries.This work provides a robust solution for personalized travel planning,capable of generating novel and compelling routes for both new and existing users by learning from collective travel intelligence.展开更多
Conversational recommender systems(CRSs)focus on refining preferences and providing personalized recommendations through natural language interactions and dialogue history.Large language models(LLMs)have shown outstan...Conversational recommender systems(CRSs)focus on refining preferences and providing personalized recommendations through natural language interactions and dialogue history.Large language models(LLMs)have shown outstanding performance across various domains,thereby prompting researchers to investigate their applicability in recommendation systems.However,due to the lack of task-specific knowledge and an inefficient feature extraction process,LLMs still have suboptimal performance in recommendation tasks.Therefore,external knowledge sources,such as knowledge graphs(KGs)and knowledge bases(KBs),are often introduced to address the issue of data sparsity.Compared to KGs,KBs possess higher retrieval efficiency,making them more suitable for scenarios where LLMs serve as recommenders.To this end,we introduce a novel framework integrating LLMs with KBs for enhanced retrieval generation,namely LLMKB.LLMKB initially leverages structured knowledge to create mapping dictionaries,extracting entity-relation information from heterogeneous knowledge to construct KBs.Then,LLMKB achieves the embedding calibration between user information representations and documents in KBs through retrieval model fine-tuning.Finally,LLMKB employs retrievalaugmented generation to produce recommendations based on fused text inputs,followed by post-processing.Experiment results on two public CRS datasets demonstrate the effectiveness of our framework.Our code is publicly available at the link:https://anonymous.4open.science/r/LLMKB-6FD0.展开更多
Content-based filtering E-commerce recommender system was discussed fully in this paper. Users＇ unique features can be explored by means of vector space model firstly. Then based on the qualitative value of products ...Content-based filtering E-commerce recommender system was discussed fully in this paper. Users＇ unique features can be explored by means of vector space model firstly. Then based on the qualitative value of products informa tion, the recommender lists were obtained. Since the system can adapt to the users＇ feedback automatically, its performance were enhanced comprehensively. Finally the evaluation of the system and the experimental results were presented.展开更多
A well-performed recommender system for an e-commerce web site can help customers easily find favorite items and then increase the turnover of merchants, hence it is important for both customers and merchants. In most...A well-performed recommender system for an e-commerce web site can help customers easily find favorite items and then increase the turnover of merchants, hence it is important for both customers and merchants. In most of the existing recommender systems, only the purchase information is utilized data and the navigational and behavioral data are seldom concerned. In this paper, we design a novel recommender system for comprehensive online shopping sites. In the proposed recommender system, the navigational and behavioral data, such as access, click, read, and purchase information of a customer, are utilized to calculate the preference degree to each item; then items with larger preference degrees are recommended to the customer. The proposed method has several innovations and two of them are more remarkable: one is that nonexpendable items are distinguished from expendable ones and handled by a different way; another is that the interest shifting of customers are considered. Lastly, we structure an example to show the operation procedure and the performance of the proposed recommender system. The results show that the proposed recommender method with considering interest shifting is superior to Kim et al(2011) method and the method without considering interest shifting.展开更多
In this work, Kendall correlation based collaborative filtering algorithms for the recommender systems are proposed. The Kendall correlation method is used to measure the correlation amongst users by means of consider...In this work, Kendall correlation based collaborative filtering algorithms for the recommender systems are proposed. The Kendall correlation method is used to measure the correlation amongst users by means of considering the relative order of the users＇ ratings. Kendall based algorithm is based upon a more general model and thus could be more widely applied in e-commerce. Another discovery of this work is that the consideration of only positive correlated neighbors in prediction, in both Pearson and Kendall algorithms, achieves higher accuracy than the consideration of all neighbors, with only a small loss of coverage.展开更多
Recommender systems are very useful for people to explore what they really need.Academic papers are important achievements for researchers and they often have a great deal of choice to submit their papers.In order to ...Recommender systems are very useful for people to explore what they really need.Academic papers are important achievements for researchers and they often have a great deal of choice to submit their papers.In order to improve the efficiency of selecting the most suitable journals for publishing their works,journal recommender systems(JRS)can automatically provide a small number of candidate journals based on key information such as the title and the abstract.However,users or journal owners may attack the system for their own purposes.In this paper,we discuss about the adversarial attacks against content-based filtering JRS.We propose both targeted attack method that makes some target journals appear more often in the system and non-targeted attack method that makes the system provide incorrect recommendations.We also conduct extensive experiments to validate the proposed methods.We hope this paper could help improve JRS by realizing the existence of such adversarial attacks.展开更多
Collaborative Filtering(CF) is a leading approach to build recommender systems which has gained considerable development and popularity. A predominant approach to CF is rating prediction recommender algorithm, aiming ...Collaborative Filtering(CF) is a leading approach to build recommender systems which has gained considerable development and popularity. A predominant approach to CF is rating prediction recommender algorithm, aiming to predict a user's rating for those items which were not rated yet by the user. However, with the increasing number of items and users, thedata is sparse.It is difficult to detectlatent closely relation among the items or users for predicting the user behaviors. In this paper,we enhance the rating prediction approach leading to substantial improvement of prediction accuracy by categorizing according to the genres of movies. Then the probabilities that users are interested in the genres are computed to integrate the prediction of each genre cluster. A novel probabilistic approach based on the sentiment analysis of the user reviews is also proposed to give intuitional explanations of why an item is recommended.To test the novel recommendation approach, a new corpus of user reviews on movies obtained from the Internet Movies Database(IMDB) has been generated. Experimental results show that the proposed framework is effective and achieves a better prediction performance.展开更多
With the rapid development of social network in recent years, a huge number of social information has been produced. As traditional recommender systems often face data sparsity and cold-start problem, the use of socia...With the rapid development of social network in recent years, a huge number of social information has been produced. As traditional recommender systems often face data sparsity and cold-start problem, the use of social information has attracted many researchers' attention to improve the prediction accuracy of recommender systems. Social trust and social relation have been proven useful to improve the performance of recommendation. Based on the classic collaborative filtering technique, we propose a PCCTTF recommender method that takes the rating time of users, social trust among users, and item tags into consideration, then do the item recommending. Experimental results show that the PCCTTF method has better prediction accuracy than classical collaborative filtering technique and the state-of-the-art recommender methods, and can also effectively alleviate data sparsity and cold-start problem. Furthermore, the PCCTTF method has better performance than all the compared methods while counting against shilling attacks.展开更多
A new recommendation method was presented based on memetic algorithm-based clustering. The proposed method was tested on four highly sparse real-world datasets. Its recommendation performance is evaluated and compared...A new recommendation method was presented based on memetic algorithm-based clustering. The proposed method was tested on four highly sparse real-world datasets. Its recommendation performance is evaluated and compared with that of the frequency-based, user-based, item-based, k-means clustering-based, and genetic algorithm-based methods in terms of precision, recall, and F1 score. The results show that the proposed method yields better performance under the new user cold-start problem when each of new active users selects only one or two items into the basket. The average F1 scores on all four datasets are improved by 225.0%, 61.6%, 54.6%, 49.3%, 28.8%, and 6.3% over the frequency-based, user-based, item-based, k-means clustering-based, and two genetic algorithm-based methods, respectively.展开更多
Memory-based collaborative recommender system (CRS) computes the similarity between users based on their declared ratings. However, not all ratings are of the same importance to the user. The set of ratings each user ...Memory-based collaborative recommender system (CRS) computes the similarity between users based on their declared ratings. However, not all ratings are of the same importance to the user. The set of ratings each user weights highly differs from user to user according to his mood and taste. This is usually reflected in the user’s rating scale. Accordingly, many efforts have been done to introduce weights to the similarity measures of CRSs. This paper proposes fuzzy weightings for the most common similarity measures for memory-based CRSs. Fuzzy weighting can be considered as a learning mechanism for capturing the preferences of users for ratings. Comparing with genetic algorithm learning, fuzzy weighting is fast, effective and does not require any more space. Moreover, fuzzy weightings based on the rating deviations from the user’s mean of ratings take into account the different rating scales of different users. The experimental results show that fuzzy weightings obviously improve the CRSs performance to a good extent.展开更多
Many datasets in E-commerce have rich information about items and users who purchase or rate them. This information can enable advanced machine learning algorithms to extract and assign user sentiments to various aspe...Many datasets in E-commerce have rich information about items and users who purchase or rate them. This information can enable advanced machine learning algorithms to extract and assign user sentiments to various aspects of the items thus leading to more sophisticated and justifiable recommendations. However, most Collaborative Filtering (CF) techniques rely mainly on the overall preferences of users toward items only. And there is lack of conceptual and computational framework that enables an understandable aspect-based AI approach to recommending items to users. In this paper, we propose concepts and computational tools that can sharpen the logic of recommendations and that rely on users’ sentiments along various aspects of items. These concepts include: The sentiment of a user towards a specific aspect of a specific item, the emphasis that a given user places on a specific aspect in general, the popularity and controversy of an aspect among groups of users, clusters of users emphasizing a given aspect, clusters of items that are popular among a group of users and so forth. The framework introduced in this study is developed in terms of user emphasis, aspect popularity, aspect controversy, and users and items similarity. Towards this end, we introduce the Aspect-Based Collaborative Filtering Toolbox (ABCFT), where the tools are all developed based on the three-index sentiment tensor with the indices being the user, item, and aspect. The toolbox computes solutions to the questions alluded to above. We illustrate the methodology using a hotel review dataset having around 6000 users, 400 hotels and 6 aspects.展开更多
We discuss some methods for constructing recommender systems. An important feature of the methods studied here is that we assume the availability of a description, representation, of the objects being considered for r...We discuss some methods for constructing recommender systems. An important feature of the methods studied here is that we assume the availability of a description, representation, of the objects being considered for recommendation. The approaches studied here differ from collaborative filtering in that we only use preferences information from the individual for whom we are providing the recommendation and make no use the preferences of other collaborators. We provide a detailed discussion of the construction of the representation schema used. We consider two sources of information about the users preferences. The first are direct statements about the type of objects the user likes. The second source of information comes from ratings of objects which the user has experienced.展开更多
The rapid growth in software demand incentivizes software development organizations to develop exclusive software for their customers worldwide.This problem is addressed by the software development industry by softwar...The rapid growth in software demand incentivizes software development organizations to develop exclusive software for their customers worldwide.This problem is addressed by the software development industry by software product line(SPL)practices that employ feature models.However,optimal feature selection based on user requirements is a challenging task.Thus,there is a requirement to resolve the challenges of software development,to increase satisfaction and maintain high product quality,for massive customer needs within limited resources.In this work,we propose a recommender system for the development team and clients to increase productivity and quality by utilizing historical information and prior experiences of similar developers and clients.The proposed system recommends features with their estimated cost concerning new software requirements,from all over the globe according to similar developers’and clients’needs and preferences.The system guides and facilitates the development team by suggesting a list of features,code snippets,libraries,cheat sheets of programming languages,and coding references from a cloud-based knowledge management repository.Similarly,a list of features is suggested to the client according to their needs and preferences.The experimental results revealed that the proposed recommender system is feasible and effective,providing better recommendations to developers and clients.It provides proper and reasonably well-estimated costs to perform development tasks effectively as well as increase the client’s satisfaction level.The results indicate that there is an increase in productivity,performance,and quality of products and a reduction in effort,complexity,and system failure.Therefore,our proposed system facilitates developers and clients during development by providing better recommendations in terms of solutions and anticipated costs.Thus,the increase in productivity and satisfaction level maximizes the benefits and usability of SPL in the modern era of technology.展开更多
文摘Selecting appropriate tourist attractions to visit in real time is an important problem for travellers.Since recommenders proactively suggest items based on user preference,they are a promising solution for this problem.Travellers visit tourist attractions sequentially by considering multiple attributes at the same time.Therefore,it is desirable to consider this when developing recommenders for tourist attractions.Using GRU4REC,we proposed RNN-based sequence-aware recommenders(RNN-SARs)that use multiple sequence datasets for training the recommended model,named multi-RNN-SARs.We proposed two types of multi-RNN-SARs-concatenate-RNN-SARs and parallel-RNN-SARs.In order to evaluate multi-RNN-SARs,we compared hit rate(HR)and mean reciprocal rank(MRR)of the item-based collaborative filtering recommender(item-CFR),RNN-SAR with the single-sequence dataset(basic-RNN-SAR),multi-RNN-SARs and the state-of-the-art SARs using a real-world travel dataset.Our research shows that multi-RNN-SARs have significantly higher performances compared to item-CFR.Not all multi-RNNSARs outperform basic-RNN-SAR but the best multi-RNN-SAR achieves comparable performance to that of the state-of-the-art algorithms.These results highlight the importance of using multiple sequence datasets in RNN-SARs and the importance of choosing appropriate sequence datasets and learning methods for implementing multi-RNN-SARs in practice.
文摘A recommender system is a tool designed to suggest relevant items to users based on their preferences and behaviors.Collaborative filtering,a popular technique within recommender systems,predicts user interests by analyzing patterns in interactions and similarities between users,leveraging past behavior data to make personalized recommendations.Despite its popularity,collaborative filtering faces notable challenges,and one of them is the issue of grey-sheep users who have unusual tastes in the system.Surprisingly,existing research has not extensively explored outlier detection techniques to address the grey-sheep problem.To fill this research gap,this study conducts a comprehensive comparison of 12 outlier detectionmethods(such as LOF,ABOD,HBOS,etc.)and introduces innovative user representations aimed at improving the identification of outliers within recommender systems.More specifically,we proposed and examined three types of user representations:1)the distribution statistics of user-user similarities,where similarities were calculated based on users’rating vectors;2)the distribution statistics of user-user similarities,but with similarities derived from users represented by latent factors;and 3)latent-factor vector representations.Our experiments on the Movie Lens and Yahoo!Movie datasets demonstrate that user representations based on latent-factor vectors consistently facilitate the identification of more grey-sheep users when applying outlier detection methods.
文摘This study introduces an advanced recommender system for technology enhanced learning(TEL)that synergizes neural collaborative filtering,sentiment analysis,and an adaptive learning rate to address the limitations of traditional TEL systems.Recognizing the critical gap in existing approaches—primarily their neglect of user emotional feedback and static learning paths—our model innovatively incorporates sentiment analysis to capture and respond to nuanced emotional feedback from users.Utilizing bidirectional encoder representations from Transformers for sentiment analysis,our system not only understands but also respects user privacy by processing feedback without revealing sensitive information.The adaptive learning rate,inspired by AdaGrad,allows our model to adjust its learning trajectory based on the sentiment scores associated with user feedback,ensuring a dynamic response to both positive and negative sentiments.This dual approach enhances the system’s adapt-ability to changing user preferences and improves its contentment understanding.Our methodology involves a comprehensive analysis of both the content of learning materials and the behaviors and preferences of learners,facilitating a more personalized learning experience.By dynamically adjusting recommendations based on real-time user data and behavioral analysis,our system leverages the collective insights of similar users and rele-vant content.We validated our approach against three datasets-MovieLens,Amazon,and a proprietary TEL dataset—and saw significant improvements in recommendation precision,F-score,and mean absolute error.The results indicate the potential of integrating sentiment analysis and adaptive learning rates into TEL recommender systems,marking a step forward in developing more responsive and user-centric educational technologies.This study paves the way for future advancements in TEL systems,emphasizing the importance of emotional intelli-gence and adaptability in enhancing the learning experience.
基金funded by the National Nature Sciences Foundation of China with Grant No.42250410321。
文摘Music recommendation systems are essential due to the vast amount of music available on streaming platforms,which can overwhelm users trying to find new tracks that match their preferences.These systems analyze users’emotional responses,listening habits,and personal preferences to provide personalized suggestions.A significant challenge they face is the“cold start”problem,where new users have no past interactions to guide recommendations.To improve user experience,these systems aimto effectively recommendmusic even to such users by considering their listening behavior and music popularity.This paper introduces a novel music recommendation system that combines order clustering and a convolutional neural network,utilizing user comments and rankings as input.Initially,the system organizes users into clusters based on semantic similarity,followed by the utilization of their rating similarities as input for the convolutional neural network.This network then predicts ratings for unreviewed music by users.Additionally,the system analyses user music listening behaviour and music popularity.Music popularity can help to address cold start users as well.Finally,the proposed method recommends unreviewed music based on predicted high rankings and popularity,taking into account each user’s music listening habits.The proposed method combines predicted high rankings and popularity by first selecting popular unreviewedmusic that themodel predicts to have the highest ratings for each user.Among these,the most popular tracks are prioritized,defined by metrics such as frequency of listening across users.The number of recommended tracks is aligned with each user’s typical listening rate.The experimental findings demonstrate that the new method outperformed other classification techniques and prior recommendation systems,yielding a mean absolute error(MAE)rate and rootmean square error(RMSE)rate of approximately 0.0017,a hit rate of 82.45%,an average normalized discounted cumulative gain(nDCG)of 82.3%,and a prediction accuracy of new ratings at 99.388%.
文摘The performance of deep recommendation models degrades significantly under data poisoning attacks.While adversarial training methods such as Vulnerability-Aware Training(VAT)enhance robustness by injecting perturbations into embeddings,they remain limited by coarse-grained noise and a static defense strategy,leaving models susceptible to adaptive attacks.This study proposes a novel framework,Self-Purification Data Sanitization(SPD),which integrates vulnerability-aware adversarial training with dynamic label correction.Specifically,SPD first identifies high-risk users through a fragility scoring mechanism,then applies self-purification by replacing suspicious interactions with model-predicted high-confidence labels during training.This closed-loop process continuously sanitizes the training data and breaks the protection ceiling of conventional adversarial training.Experiments demonstrate that SPD significantly improves the robustness of both Matrix Factorization(MF)and LightGCN models against various poisoning attacks.We show that SPD effectively suppresses malicious gradient propagation and maintains recommendation accuracy.Evaluations on Gowalla and Yelp2018 confirmthat SPD-trainedmodels withstandmultiple attack strategies—including Random,Bandwagon,DP,and Rev attacks—while preserving performance.
基金funded by Soonchunhyang University,Grant Number 20250029。
文摘Recommendation systems have become indispensable for providing tailored suggestions and capturing evolving user preferences based on interaction histories.The collaborative filtering(CF)model,which depends exclusively on user-item interactions,commonly encounters challenges,including the cold-start problem and an inability to effectively capture the sequential and temporal characteristics of user behavior.This paper introduces a personalized recommendation system that combines deep learning techniques with Bayesian Personalized Ranking(BPR)optimization to address these limitations.With the strong support of Long Short-Term Memory(LSTM)networks,we apply it to identify sequential dependencies of user behavior and then incorporate an attention mechanism to improve the prioritization of relevant items,thereby enhancing recommendations based on the hybrid feedback of the user and its interaction patterns.The proposed system is empirically evaluated using publicly available datasets from movie and music,and we evaluate the performance against standard recommendation models,including Popularity,BPR,ItemKNN,FPMC,LightGCN,GRU4Rec,NARM,SASRec,and BERT4Rec.The results demonstrate that our proposed framework consistently achieves high outcomes in terms of HitRate,NDCG,MRR,and Precision at K=100,with scores of(0.6763,0.1892,0.0796,0.0068)on MovieLens-100K,(0.6826,0.1920,0.0813,0.0068)on MovieLens-1M,and(0.7937,0.3701,0.2756,0.0078)on Last.fm.The results show an average improvement of around 15%across all metrics compared to existing sequence models,proving that our framework ranks and recommends items more accurately.
基金supported by the National Key R&D Program of China[2022YFF0902703]the State Administration for Market Regulation Science and Technology Plan Project(2024MK033).
文摘Recommendation systems are key to boosting user engagement,satisfaction,and retention,particularly on media platforms where personalized content is vital.Sequential recommendation systems learn from user-item interactions to predict future items of interest.However,many current methods rely on unique user and item IDs,limiting their ability to represent users and items effectively,especially in zero-shot learning scenarios where training data is scarce.With the rapid development of Large Language Models(LLMs),researchers are exploring their potential to enhance recommendation systems.However,there is a semantic gap between the linguistic semantics of LLMs and the collaborative semantics of recommendation systems,where items are typically indexed by IDs.Moreover,most research focuses on item representations,neglecting personalized user modeling.To address these issues,we propose a sequential recommendation framework using LLMs,called CIT-Rec,a model that integrates Collaborative semantics for user representation and Image and Text information for item representation to enhance Recommendations.Specifically,by aligning intuitive image information with text containing semantic features,we can more accurately represent items,improving item representation quality.We focus not only on item representations but also on user representations.To more precisely capture users’personalized preferences,we use traditional sequential recommendation models to train on users’historical interaction data,effectively capturing behavioral patterns.Finally,by combining LLMs and traditional sequential recommendation models,we allow the LLM to understand linguistic semantics while capturing collaborative semantics.Extensive evaluations on real-world datasets show that our model outperforms baseline methods,effectively combining user interaction history with item visual and textual modalities to provide personalized recommendations.
基金supported by the Chung-Ang University Research Grants in 2023.Alsothe work is supported by the ELLIIT Excellence Center at Linköping–Lund in Information Technology in Sweden.
文摘Recommending personalized travel routes from sparse,implicit feedback poses a significant challenge,as conventional systems often struggle with information overload and fail to capture the complex,sequential nature of user preferences.To address this,we propose a Conditional Generative Adversarial Network(CGAN)that generates diverse and highly relevant itineraries.Our approach begins by constructing a conditional vector that encapsulates a user’s profile.This vector uniquely fuses embeddings from a Heterogeneous Information Network(HIN)to model complex user-place-route relationships,a Recurrent Neural Network(RNN)to capture sequential path dynamics,and Neural Collaborative Filtering(NCF)to incorporate collaborative signals from the wider user base.This comprehensive condition,further enhanced with features representing user interaction confidence and uncertainty,steers a CGAN stabilized by spectral normalization to generate high-fidelity latent route representations,effectively mitigating the data sparsity problem.Recommendations are then formulated using an Anchor-and-Expand algorithm,which selects relevant starting Points of Interest(POI)based on user history,then expands routes through latent similarity matching and geographic coherence optimization,culminating in Traveling Salesman Problem(TSP)-based route optimization for practical travel distances.Experiments on a real-world check-in dataset validate our model’s unique generative capability,achieving F1 scores ranging from 0.163 to 0.305,and near-zero pairs−F1 scores between 0.002 and 0.022.These results confirm the model’s success in generating novel travel routes by recommending new locations and sequences rather than replicating users’past itineraries.This work provides a robust solution for personalized travel planning,capable of generating novel and compelling routes for both new and existing users by learning from collective travel intelligence.
文摘Conversational recommender systems(CRSs)focus on refining preferences and providing personalized recommendations through natural language interactions and dialogue history.Large language models(LLMs)have shown outstanding performance across various domains,thereby prompting researchers to investigate their applicability in recommendation systems.However,due to the lack of task-specific knowledge and an inefficient feature extraction process,LLMs still have suboptimal performance in recommendation tasks.Therefore,external knowledge sources,such as knowledge graphs(KGs)and knowledge bases(KBs),are often introduced to address the issue of data sparsity.Compared to KGs,KBs possess higher retrieval efficiency,making them more suitable for scenarios where LLMs serve as recommenders.To this end,we introduce a novel framework integrating LLMs with KBs for enhanced retrieval generation,namely LLMKB.LLMKB initially leverages structured knowledge to create mapping dictionaries,extracting entity-relation information from heterogeneous knowledge to construct KBs.Then,LLMKB achieves the embedding calibration between user information representations and documents in KBs through retrieval model fine-tuning.Finally,LLMKB employs retrievalaugmented generation to produce recommendations based on fused text inputs,followed by post-processing.Experiment results on two public CRS datasets demonstrate the effectiveness of our framework.Our code is publicly available at the link:https://anonymous.4open.science/r/LLMKB-6FD0.
基金Supported bythe Hunan Teaching Reformand Re-search Project of Colleges and Universities (2003-B72) the HunanBoard of Review on Philosophic and Social Scientific Pay-off Project(0406035) the Hunan Soft Science Research Project(04ZH6005)
文摘Content-based filtering E-commerce recommender system was discussed fully in this paper. Users＇ unique features can be explored by means of vector space model firstly. Then based on the qualitative value of products informa tion, the recommender lists were obtained. Since the system can adapt to the users＇ feedback automatically, its performance were enhanced comprehensively. Finally the evaluation of the system and the experimental results were presented.
基金supported by theNational High-Tech R&D Program (863 Program) No. 2015AA01A705the National Natural Science Foundation of China under Grant No. 61572072+2 种基金the National Science and Technology Major Project No. 2015ZX03001041Fundamental Research Funds for the Central Universities No. FRF-TP-15-027A3Yunnan Provincial Department of Education Foundation Project (No. 2014Y087)
文摘A well-performed recommender system for an e-commerce web site can help customers easily find favorite items and then increase the turnover of merchants, hence it is important for both customers and merchants. In most of the existing recommender systems, only the purchase information is utilized data and the navigational and behavioral data are seldom concerned. In this paper, we design a novel recommender system for comprehensive online shopping sites. In the proposed recommender system, the navigational and behavioral data, such as access, click, read, and purchase information of a customer, are utilized to calculate the preference degree to each item; then items with larger preference degrees are recommended to the customer. The proposed method has several innovations and two of them are more remarkable: one is that nonexpendable items are distinguished from expendable ones and handled by a different way; another is that the interest shifting of customers are considered. Lastly, we structure an example to show the operation procedure and the performance of the proposed recommender system. The results show that the proposed recommender method with considering interest shifting is superior to Kim et al(2011) method and the method without considering interest shifting.
基金Supported by the National Natural Science Foun-dation of China (60573095)
文摘In this work, Kendall correlation based collaborative filtering algorithms for the recommender systems are proposed. The Kendall correlation method is used to measure the correlation amongst users by means of considering the relative order of the users＇ ratings. Kendall based algorithm is based upon a more general model and thus could be more widely applied in e-commerce. Another discovery of this work is that the consideration of only positive correlated neighbors in prediction, in both Pearson and Kendall algorithms, achieves higher accuracy than the consideration of all neighbors, with only a small loss of coverage.
基金This work is supported by the National Natural Science Foundation of China under Grant Nos.U1636215,61902082the Guangdong Key R&D Program of China 2019B010136003Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme(2019).
文摘Recommender systems are very useful for people to explore what they really need.Academic papers are important achievements for researchers and they often have a great deal of choice to submit their papers.In order to improve the efficiency of selecting the most suitable journals for publishing their works,journal recommender systems(JRS)can automatically provide a small number of candidate journals based on key information such as the title and the abstract.However,users or journal owners may attack the system for their own purposes.In this paper,we discuss about the adversarial attacks against content-based filtering JRS.We propose both targeted attack method that makes some target journals appear more often in the system and non-targeted attack method that makes the system provide incorrect recommendations.We also conduct extensive experiments to validate the proposed methods.We hope this paper could help improve JRS by realizing the existence of such adversarial attacks.
基金supported in part by National Science Foundation of China under Grants No.61303105 and 61402304the Humanity&Social Science general project of Ministry of Education under Grants No.14YJAZH046+2 种基金the Beijing Natural Science Foundation under Grants No.4154065the Beijing Educational Committee Science and Technology Development Planned under Grants No.KM201410028017Academic Degree Graduate Courses group projects
文摘Collaborative Filtering(CF) is a leading approach to build recommender systems which has gained considerable development and popularity. A predominant approach to CF is rating prediction recommender algorithm, aiming to predict a user's rating for those items which were not rated yet by the user. However, with the increasing number of items and users, thedata is sparse.It is difficult to detectlatent closely relation among the items or users for predicting the user behaviors. In this paper,we enhance the rating prediction approach leading to substantial improvement of prediction accuracy by categorizing according to the genres of movies. Then the probabilities that users are interested in the genres are computed to integrate the prediction of each genre cluster. A novel probabilistic approach based on the sentiment analysis of the user reviews is also proposed to give intuitional explanations of why an item is recommended.To test the novel recommendation approach, a new corpus of user reviews on movies obtained from the Internet Movies Database(IMDB) has been generated. Experimental results show that the proposed framework is effective and achieves a better prediction performance.
基金Supported by the National Natural Science Foundation of China(71662014,61602219,71861013)。
文摘With the rapid development of social network in recent years, a huge number of social information has been produced. As traditional recommender systems often face data sparsity and cold-start problem, the use of social information has attracted many researchers' attention to improve the prediction accuracy of recommender systems. Social trust and social relation have been proven useful to improve the performance of recommendation. Based on the classic collaborative filtering technique, we propose a PCCTTF recommender method that takes the rating time of users, social trust among users, and item tags into consideration, then do the item recommending. Experimental results show that the PCCTTF method has better prediction accuracy than classical collaborative filtering technique and the state-of-the-art recommender methods, and can also effectively alleviate data sparsity and cold-start problem. Furthermore, the PCCTTF method has better performance than all the compared methods while counting against shilling attacks.
基金supporting by grant fund under the Strategic Scholarships for Frontier Research Network for the PhD Program Thai Doctoral degree
文摘A new recommendation method was presented based on memetic algorithm-based clustering. The proposed method was tested on four highly sparse real-world datasets. Its recommendation performance is evaluated and compared with that of the frequency-based, user-based, item-based, k-means clustering-based, and genetic algorithm-based methods in terms of precision, recall, and F1 score. The results show that the proposed method yields better performance under the new user cold-start problem when each of new active users selects only one or two items into the basket. The average F1 scores on all four datasets are improved by 225.0%, 61.6%, 54.6%, 49.3%, 28.8%, and 6.3% over the frequency-based, user-based, item-based, k-means clustering-based, and two genetic algorithm-based methods, respectively.
文摘Memory-based collaborative recommender system (CRS) computes the similarity between users based on their declared ratings. However, not all ratings are of the same importance to the user. The set of ratings each user weights highly differs from user to user according to his mood and taste. This is usually reflected in the user’s rating scale. Accordingly, many efforts have been done to introduce weights to the similarity measures of CRSs. This paper proposes fuzzy weightings for the most common similarity measures for memory-based CRSs. Fuzzy weighting can be considered as a learning mechanism for capturing the preferences of users for ratings. Comparing with genetic algorithm learning, fuzzy weighting is fast, effective and does not require any more space. Moreover, fuzzy weightings based on the rating deviations from the user’s mean of ratings take into account the different rating scales of different users. The experimental results show that fuzzy weightings obviously improve the CRSs performance to a good extent.
文摘Many datasets in E-commerce have rich information about items and users who purchase or rate them. This information can enable advanced machine learning algorithms to extract and assign user sentiments to various aspects of the items thus leading to more sophisticated and justifiable recommendations. However, most Collaborative Filtering (CF) techniques rely mainly on the overall preferences of users toward items only. And there is lack of conceptual and computational framework that enables an understandable aspect-based AI approach to recommending items to users. In this paper, we propose concepts and computational tools that can sharpen the logic of recommendations and that rely on users’ sentiments along various aspects of items. These concepts include: The sentiment of a user towards a specific aspect of a specific item, the emphasis that a given user places on a specific aspect in general, the popularity and controversy of an aspect among groups of users, clusters of users emphasizing a given aspect, clusters of items that are popular among a group of users and so forth. The framework introduced in this study is developed in terms of user emphasis, aspect popularity, aspect controversy, and users and items similarity. Towards this end, we introduce the Aspect-Based Collaborative Filtering Toolbox (ABCFT), where the tools are all developed based on the three-index sentiment tensor with the indices being the user, item, and aspect. The toolbox computes solutions to the questions alluded to above. We illustrate the methodology using a hotel review dataset having around 6000 users, 400 hotels and 6 aspects.
文摘We discuss some methods for constructing recommender systems. An important feature of the methods studied here is that we assume the availability of a description, representation, of the objects being considered for recommendation. The approaches studied here differ from collaborative filtering in that we only use preferences information from the individual for whom we are providing the recommendation and make no use the preferences of other collaborators. We provide a detailed discussion of the construction of the representation schema used. We consider two sources of information about the users preferences. The first are direct statements about the type of objects the user likes. The second source of information comes from ratings of objects which the user has experienced.
基金supported by the National Natural Science Foundation of China(Grant Number:61672080,Sponsored Authors:Yang S.,Sponsors’Websites:http://www.nsfc.gov.cn/english/site_1/index.html).
文摘The rapid growth in software demand incentivizes software development organizations to develop exclusive software for their customers worldwide.This problem is addressed by the software development industry by software product line(SPL)practices that employ feature models.However,optimal feature selection based on user requirements is a challenging task.Thus,there is a requirement to resolve the challenges of software development,to increase satisfaction and maintain high product quality,for massive customer needs within limited resources.In this work,we propose a recommender system for the development team and clients to increase productivity and quality by utilizing historical information and prior experiences of similar developers and clients.The proposed system recommends features with their estimated cost concerning new software requirements,from all over the globe according to similar developers’and clients’needs and preferences.The system guides and facilitates the development team by suggesting a list of features,code snippets,libraries,cheat sheets of programming languages,and coding references from a cloud-based knowledge management repository.Similarly,a list of features is suggested to the client according to their needs and preferences.The experimental results revealed that the proposed recommender system is feasible and effective,providing better recommendations to developers and clients.It provides proper and reasonably well-estimated costs to perform development tasks effectively as well as increase the client’s satisfaction level.The results indicate that there is an increase in productivity,performance,and quality of products and a reduction in effort,complexity,and system failure.Therefore,our proposed system facilitates developers and clients during development by providing better recommendations in terms of solutions and anticipated costs.Thus,the increase in productivity and satisfaction level maximizes the benefits and usability of SPL in the modern era of technology.
