Collaborative Filtering Problem Research Articles

A hybrid similarity model for mitigating the cold-start problem of collaborative filtering in sparse data

Similarity is a vital component for neighborhood-based collaborative filtering (CF). To improve the quality of recommendation, many similarity methods have been proposed and analyzed in recent decades. However, nearly all traditional similarity methods and many advanced similarity methods only utilize corated items among users to compute their similarity, which provides limited information in cold-start/sparse scenarios and yields misleading results. In addition, although a few advanced hybrid similarity models consider items beyond corated items, which can partly mitigate the above limitation, they still have drawbacks, such as disregarding penalizing noncorated items that have many disadvantages. In this paper, we explore a new robust hybrid similarity model, namely Wasserstein distance-based CF (WCF) model, for mitigating the cold-start problem of CF in sparse data. Specifically, we measure item similarity via the Wasserstein distance, which can help circumvent the drawbacks in the Bhattacharyya coefficient and KL divergence that are used in the literature, and is thus more robust in a cold-start/sparse scenario. Besides, we further design a new multiplicative user similarity formula which identifies all noncorated items as a whole to prioritize the importance of corated items and impair the negative effects of noncorated items, which will also play an important role in a cold-start/sparse scenario. In addition, we also propose two novel heuristic similarity factors to impair the negative effects of popular users and items as supplements. We conduct extensive experiments on five real-world benchmark recommendation datasets to test WCF. The experimental results show the superiority of WCF over other existing similarity methods in cold-start/sparse scenarios.

SimDCL: dropout-based simple graph contrastive learning for recommendation

Representation learning of users and items is the core of recommendation, and benefited from the development of graph neural network (GNN), graph collaborative filtering (GCF) for capturing higher order connectivity has been successful in the recommendation domain. Nevertheless, the matrix sparsity problem in collaborative filtering and the tendency of higher order embeddings to smooth in GNN limit further performance improvements. Contrastive learning (CL) was introduced into GCF and alleviated these problems to some extent. However, existing methods usually require graph perturbation to construct augmented views or design complex CL tasks, which limits the further development of CL-based methods in the recommendation. We propose a simple CL framework that does not require graph augmentation, but is based on dropout techniques to generate contrastive views to address the aforementioned problem. Specifically, we first added dropout operation to the GNN computation, and then fed the same batch of samples twice into the network for computation. Using the randomness of dropout, a pair of views with random noise was obtained, and maximizing the similarity of the view pairs is set as an auxiliary task to complement the recommendation. In addition, we made a simple modification to the computation of the GNN to alleviate the information loss due to embedding smoothing by means of cross-layer connected graph convolution computation. We named our proposed method as Simple Contrastive Learning Graph Neural Network based on dropout (SimDCL). Extensive experiments on five public datasets demonstrate the effectiveness of the proposed SimDCL, especially on the Amazon Books and Ta-Feng datasets, where our approach achieves 44% and 43% performance gains compared to baseline.

KEMIM: Knowledge-Enhanced User Multi-Interest Modeling for Recommender Systems

Researchers typically leverage side information, such as social networks or the knowledge graph, to overcome the sparsity and cold start problem in collaborative filtering. To tackle the limitations of existing user interest modeling, we propose a knowledge-enhanced user multi-interest modeling for recommender systems (KEMIM). First, we utilize the user-item historical interaction as the knowledge graph’s head entity to create a user’s explicit interests and leverage the relationship path to expand the user’s potential interests through connections in the knowledge graph. Second, considering the diversity of a user’s interests, we adopt an attention mechanism to learn the user’s attention to each historical interaction and each potential interest. Third, we combine the user’s attribute features with interests to solve the cold start problem effectively. With the knowledge graph’s structural data, KEMIM could describe the features of users at a fine granularity and provide explainable recommendation results to users. In this study, we conduct an in-depth empirical evaluation across three open datasets for two different recommendation tasks: Click-Through rate (CTR) prediction and Top-K recommendation. The experimental findings demonstrate that KEMIM outperforms several state-of-the-art baselines.

Towards addressing item cold-start problem in collaborative filtering by embedding agglomerative clustering and FP-growth into the recommendation system

This paper introduces a frequent pattern mining framework for recommender systems (FPRS) - a novel approach to address the items? cold-start problem. This difficulty occurs when a new item hits the system, and properly handling such a situation is one of the key success factors of any deployment. The article proposes several strategies to combine collaborative and content-based filtering methods with frequent items mining and agglomerative clustering techniques to mitigate the cold-start problem in recommender systems. The experiments evaluated the developed methods against several quality metrics on three benchmark datasets. The conducted study confirmed usefulness of FPRS in providing apt outcomes even for cold items. The presented solution can be integrated with many different approaches and further extended to make up a complete and standalone RS.

Reliable prediction in the Markov stochastic block model

We introduce the Markov Stochastic Block Model (MSBM): a growth model for community based networks where node attributes are assigned through a Markovian dynamic. We rely on HMMs’ literature to design prediction methods that are robust to local clustering errors. We focus specifically on the link prediction and collaborative filtering problems and we introduce a new model selection procedure to infer the number of hidden clusters in the network. Our approaches for reliable prediction in MSBMs are not algorithm-dependent in the sense that they can be applied using your favourite clustering tool. In this paper, we use a recent SDP method to infer the hidden communities and we provide theoretical guarantees. In particular, we identify the relevant signal-to-noise ratio (SNR) in our framework and we prove that the misclassification error decays exponentially fast with respect to this SNR.

A Learning Automata-Based Approach to Improve the Scalability of Clustering-Based Recommender Systems

One of the common techniques to reduce the scalability problem in collaborative filtering (CF)-based recommender systems is the clustering technique, which accelerates finding the nearest neighbor users in the recommendation process. Different clustering algorithms lead to improved accuracy and diversity in recommender systems. It is challenging to develop recommender systems based on clustering with decreasing scalability and simultaneously increasing accuracy. This article proposes a new clustering-based recommender system that takes into account the theoretical properties of the Learning Automata (LA) technique. The presented clustering novelty lies in the fact that employing LA technique for the user clustering in a CF-based recommender system has not been set forth so far in a way that addresses the scalability while improving accuracy. In addition, a novel similarity metric is embedded in the proposed algorithm to measure the similarity value between users. This metric is developed as like/dislike (LD) which can significantly improve the accuracy by reducing the computational cost. Extensive simulations have been performed on real-world datasets such as MovieLens and FilmTrust, which confirm the effectiveness of the proposed algorithm. In this regard, the proposed algorithm has improved the precision between 5 and 16% on average compared to the existing state-of-the-art methods such as GA-GELS, NUSCCF, NNMF, and KL-KM.

Algorithm for generating negative cases for collaborative filtering recommender

AbstractMost collaborative filtering recommenders based on deep learning utilize implicit feedback information such as likes and bookmarks to infer user preferences rather than explicit feedback such as ratings provided by users. However, collaborative filtering models utilizing implicit feedback pose a chronic problem, that is, the “one‐class” problem (also known as the one‐class collaborative filtering problem), which causes trivial solutions due to lack of negative examples. A one‐class problem free‐boosting (OCF‐B) algorithm is proposed in this study to solve the one‐class problem by considering a user's preference pattern when generating negative cases and training the model with both positive and negative cases. More specifically, the OCF‐B algorithm iteratively selects negative cases that exacerbate the loss value of the object function and replace them with other negative cases when extracting negative cases from unknown cells of the user–item interaction matrix. In an experiment using four datasets, the results show that negative cases selected meticulously via the OCF‐B algorithm improve the prediction performance not only for negative cases, but also for positive cases. In addition, the prediction performance was better than that of the existing zero‐injection method.

Addressing the Cold-Start Problem in Collaborative Filtering Through Positive-Unlabeled Learning and Multi-Target Prediction

The cold-start problem is one of the main challenges in recommender systems and specifically in collaborative filtering methods. Such methods, albeit effective, typically can not handle new items or users that do not have any prior interaction activity in the system. In this paper, we propose a novel two-step approach to address the cold-start problem. First, we view the user-item interactions in a positive unlabeled (PU) learning setting and reconstruct the interaction matrix between users and warm items, detecting missing links and recommending warm items to existing users. Second, an inductive multi-target regressor is trained on this reconstructed interaction matrix and subsequently predicts interactions for new items that enter the system. To the best of our knowledge, this is the first time that such a two-step PU learning method is proposed to address the cold-start problem in recommender systems. To evaluate the proposed approach, we employed four benchmark datasets from movie and news recommendation domains with explicit and implicit feedback. We compared our method against three other competitor approaches that address the cold-start problem and showed that our proposed method significantly outperforms them, achieving in a case an increase of 16.9% in terms of NDCG.

Addressing Cold Start Problem in Collaborative Filtering using Demographic Data with Entropy-based Methodology

Collaborative filtering based recommender systems (RS) are faced with cold start problem. This problem arises when the RS does not have enough information or opinion about a person or about a product and therefore cannot make recommendation for such person. In this paper, the demographic data of the user such as age, gender, and occupation are utilized as additional sources together with existing users’ rating to tackle the cold start problem by employing the entropy-based methodology to determine the degree of predictability. Experimental results on MovieLens dataset showed that the proposed method gives higher accuracy than other existing demographic based methods. Keywords— Cold Start, Collaborative Filtering, Entropy, Demographic Approach, Recommender Systems

Transfer Learning in Collaborative Filtering with Uncertain Ratings

To solve the sparsity problem in collaborative filtering, researchers have introduced transfer learning as a viable approach to make use of auxiliary data. Most previous transfer learning works in collaborative filtering have focused on exploiting point-wise ratings such as numerical ratings, stars, or binary ratings of likes/dislikes. However, in many real-world recommender systems, many users may be unwilling or unlikely to rate items with precision.In contrast, practitioners can turn to various non-preference data to estimate a range or rating distribution of a user's preference on an item.Such a range or rating distribution is called an uncertain rating since it represents a rating spectrum of uncertainty instead of an accurate point-wise score. In this paper, we propose an efficient transfer learning solution for collaborative filtering, known as {\em transfer by integrative factorization} (TIF), to leverage such auxiliary uncertain ratings to improve the performance of recommendation. In particular, we integrate auxiliary data of uncertain ratings as additional constraints in the target matrix factorization problem, and learn an expected rating value for each uncertain rating automatically. The advantages of our proposed approach include the efficiency and the improved effectiveness of collaborative filtering, showing that incorporating the auxiliary data of uncertain ratings can really bring a benefit. Experimental results on two movie recommendation tasks show that our TIF algorithm performs significantly better over a state-of-the-art non-transfer learning method.

A Sequential Decision Approach to Ordinal Preferences in Recommender Systems

We propose a novel sequential decision approach to modeling ordinal ratings in collaborative filtering problems. The rating process is assumed to start from the lowest level, evaluates against the latent utility at the corresponding level and moves up until a suitable ordinal level is found. Crucial to this generative process is the underlying utility random variables that govern the generation of ratings and their modelling choices. To this end, we make a novel use of the generalised extreme value distributions, which is found to be particularly suitable for our modeling tasks and at the same time, facilitate our inference and learning procedure. The proposed approach is flexible to incorporate features from both the user and the item. We evaluate the proposed framework on three well-known datasets: MovieLens, Dating Agency and Netflix. In all cases, it is demonstrated that the proposed work is competitive against state-of-the-art collaborative filtering methods.

Personalized Landmark Recommendation Based on Geotags from Photo Sharing Sites

Geotagged photos of users on social media sites provide abundant location-based data, which can be exploited for various location-based services, such as travel recommendation. In this paper, we propose a novel approach to a new application, i.e., personalized landmark recommendation based on users’ geotagged photos. We formulate the landmark recommendation task as a collaborative filtering problem, for which we propose a category-regularized matrix factorization approach that integrates both user-landmark preference and category-based landmark similarity. We collected geotagged photos from Flickr and landmark categories from Wikipedia for our experiments. Our experimental results demonstrate that the proposed approach outperforms popularity-based landmark recommendation and a basic matrix factorization approach in recommending personalized landmarks that are less visited by the population as a whole.

Variational autoencoder Bayesian matrix factorization (VABMF) for collaborative filtering

Probabilistic matrix factorization (PMF) is the most popular method among low-rank matrix approximation approaches that address the sparsity problem in collaborative filtering for recommender systems. PMF depends on the classical maximum a posteriori estimator for estimating model parameters; however, these approaches are vulnerable to overfitting because of the nature of a single point estimation that is pursued by these models. An alternative approach to PMF is a Bayesian PMF model that suggests the Markov chain Monte Carlo algorithm as a full estimation for approximate intractable posterior over model parameters. However, despite its success in increasing prediction, it has a high computational cost. To this end, we proposed a novel Bayesian deep learning-based model treatment, namely, variational autoencoder Bayesian matrix factorization (VABMF). The proposed model uses stochastic gradient variational Bayes to estimate intractable posteriors and expectation–maximization-style estimators to learn model parameters. The model was evaluated on the basis of three MovieLens datasets, namely, Ml-100k, Ml-1M, and Ml-10M. Experimental results showed that our proposed VABMF model significantly outperforms state-of-the-art RS.

An effective recommender system based on personality traits, demographics and behavior of customers in time context

PurposeImproving the performance of recommender systems (RSs) has always been a major challenge in the area of e-commerce because the systems face issues such as cold start, sparsity, scalability and interest drift that affect their performance. Despite the efforts made to solve these problems, there is still no RS that can solve or reduce all the problems simultaneously. Therefore, the purpose of this study is to provide an effective and comprehensive RS to solve or reduce all of the above issues, which uses a combination of basic customer information as well as big data techniques.Design/methodology/approachThe most important steps in the proposed RS are: (1) collecting demographic and behavioral data of customers from an e-clothing store; (2) assessing customer personality traits; (3) creating a new user-item matrix based on customer/user interest; (4) calculating the similarity between customers with efficient k-nearest neighbor (EKNN) algorithm based on locality-sensitive hashing (LSH) approach and (5) defining a new similarity function based on a combination of personality traits, demographic characteristics and time-based purchasing behavior that are the key incentives for customers' purchases.FindingsThe proposed method was compared with different baselines (matrix factorization and ensemble). The results showed that the proposed method in terms of all evaluation measures led to a significant improvement in traditional collaborative filtering (CF) performance, and with a significant difference (more than 40%), performed better than all baselines. According to the results, we find that our proposed method, which uses a combination of personality information and demographics, as well as tracking the recent interests and needs of the customer with the LSH approach, helps to improve the effectiveness of the recommendations more than the baselines. This is due to the fact that this method, which uses the above information in conjunction with the LSH technique, is more effective and more accurate in solving problems of cold start, scalability, sparsity and interest drift.Research limitations/implicationsThe research data were limited to only one e-clothing store.Practical implicationsIn order to achieve an accurate and real-time RS in e-commerce, it is essential to use a combination of customer information with efficient techniques. In this regard, according to the results of the research, the use of personality traits and demographic characteristics lead to a more accurate knowledge of customers' interests and thus better identification of similar customers. Therefore, this information should be considered as a solution to reduce the problems of cold start and sparsity. Also, a better judgment can be made about customers' interests by considering their recent purchases; therefore, in order to solve the problems of interest drifts, different weights should be assigned to purchases and launch time of products/items at different times (the more recent, the more weight). Finally, the LSH technique is used to increase the RS scalability in e-commerce. In total, a combination of personality traits, demographics and customer purchasing behavior over time with the LSH technique should be used to achieve an ideal RS. Using the RS proposed in this research, it is possible to create a comfortable and enjoyable shopping experience for customers by providing real-time recommendations that match customers' preferences and can result in an increase in the profitability of e-shops.Originality/valueIn this study, by considering a combination of personality traits, demographic characteristics and time-based purchasing behavior of customers along with the LSH technique, we were able for the first time to simultaneously solve the basic problems of CF, namely cold start, scalability, sparsity and interest drift, which led to a decrease in significant errors of recommendations and an increase in the accuracy of CF. The average errors of the recommendations provided to users based on the proposed model is only about 13%, and the accuracy and compliance of these recommendations with the interests of customers is about 92%. In addition, a 40% difference between the accuracy of the proposed method and the traditional CF method has been observed. This level of accuracy in RSs is very significant and special, which is certainly welcomed by e-business owners. This is also a new scientific finding that is very useful for programmers, users and researchers. In general, the main contributions of this research are: 1) proposing an accurate RS using personality traits, demographic characteristics and time-based purchasing behavior; 2) proposing an effective and comprehensive RS for a “clothing” online store; 3) improving the RS performance by solving the cold start issue using personality traits and demographic characteristics; 4) improving the scalability issue in RS through efficient k-nearest neighbors; 5) Mitigating the sparsity issue by using personality traits and demographic characteristics and also by densifying the user-item matrix and 6) improving the RS accuracy by solving the interest drift issue through developing a time-based user-item matrix.

A new neighbourhood formation approach for solving cold-start user problem in collaborative filtering

Collaborative filtering (CF) is the most widely accepted recommendation technique. Despite its popularity, this approach faces some major challenges like that of a cold-start user problem where a user has rated a handful of items. Due to very few ratings available for the cold-start users, their similarities with rest of the users has been questioned in the past, none have focused on their approach for neighbour identification. Whilst the traditional CF approaches select only those similar users as neighbours who have rated the item under consideration, the neighbourhood comprises of weak neighbours of the cold-start users. To address this shortcoming, our proposed approach selects neighbours with highest similarity irrespective of their availability of ratings for that item. Moreover, for the selected similar neighbours with missing ratings, an item based regression is performed to partially populate the matrix. The efficacy of the proposed neighbourhood formation approach addressing cold-start user problem is validated on two publicly available MovieLens datasets. Our approach provides superior quality of recommendations evaluated on a range of prediction and classification accuracy metrics. The results are encouraging particularly for systems having higher percentage of cold-start users which indicates the effectiveness of our approach in practical settings of new internet portals.

A new neighbourhood formation approach for solving cold-start user problem in collaborative filtering

Collaborative filtering (CF) is the most widely accepted recommendation technique. Despite its popularity, this approach faces some major challenges like that of a cold-start user problem where a user has rated a handful of items. Due to very few ratings available for the cold-start users, their similarities with rest of the users has been questioned in the past, none have focused on their approach for neighbour identification. Whilst the traditional CF approaches select only those similar users as neighbours who have rated the item under consideration, the neighbourhood comprises of weak neighbours of the cold-start users. To address this shortcoming, our proposed approach selects neighbours with highest similarity irrespective of their availability of ratings for that item. Moreover, for the selected similar neighbours with missing ratings, an item based regression is performed to partially populate the matrix. The efficacy of the proposed neighbourhood formation approach addressing cold-start user problem is validated on two publicly available MovieLens datasets. Our approach provides superior quality of recommendations evaluated on a range of prediction and classification accuracy metrics. The results are encouraging particularly for systems having higher percentage of cold-start users which indicates the effectiveness of our approach in practical settings of new internet portals.

Knowledge Graph Embedding Based Collaborative Filtering

Along with the rapidly increasing massive online data, recommender systems have been used as an effective approach for filtering useful information, which have been widely adopted in many web applications. In recent years, new techniques such as neural network and deep learning have demonstrated its effectiveness in studies of recommender systems. In this paper, we propose a novel approach for collaborative filtering with implicit feedback based on knowledge graph embedding. The basic ideal is to model the interactions between users and items as an interaction knowledge graph with a single relation, whose vector representation is learned through knowledge graph embedding. Base on the learned representation, the collaborative filtering problem is converted into link prediction in the interaction knowledge graph. KGECF, a neural network for knowledge graph embedding based collaborative filtering, is proposed based on this ideal and the RotatE knowledge graph embedding model. Experimental results on five datasets with different characteristics show that the KGECF model achieves the state-of-the-art (SOTA) performance on all datasets. And unlike other SOTA models that have obvious performance drops on AMusic and AToy datasets, our model's performance is very stable across all datasets.

A Reliable Solution for Sparsity Problem in Collaborative Filtering Using Demographic Approach

A Reliable Solution for Sparsity Problem in Collaborative Filtering Using Demographic Approach

The wisdom of the few: a provable approach

In recent years, the wisdom of the few (WOF) model has attracted substantial research interests. The WOF refers to the findings that in some collaborative prediction tasks, e.g., collaborative filtering (CF), with only the ratings from a small set of expert users, it nearly suffices to predict a much larger number of other users' unobserved ratings. In this paper, we propose a WOF algorithm for the CF problem, and prove that under some mild statistical assumptions, the algorithm can predict the users' missing ratings correctly with high probability guarantee. We also conduct CF experiments with the proposed algorithm on real datasets; the results show that our algorithm is competitive with the conventional CF algorithm.

A survey of matrix completion methods for recommendation systems

In recent years, the recommendation systems have become increasingly popular and have been used in a broad variety of applications. Here, we investigate the matrix completion techniques for the recommendation systems that are based on collaborative filtering. The collaborative filtering problem can be viewed as predicting the favorability of a user with respect to new items of commodities. When a rating matrix is constructed with users as rows, items as columns, and entries as ratings, the collaborative filtering problem can then be modeled as a matrix completion problem by filling out the unknown elements in the rating matrix. This article presents a comprehensive survey of the matrix completion methods used in recommendation systems. We focus on the mathematical models for matrix completion and the corresponding computational algorithms as well as their characteristics and potential issues. Several applications other than the traditional user-item association prediction are also discussed.